Synthesis of iron chelators. Enterobactin, enantioenterobactin, and a

Michel Meyer, Jason R. Telford, Seth M. Cohen, David J. White, Jide Xu, and Kenneth N. Raymond. Journal of the American Chemical Society 1997 119 (42)...
0 downloads 0 Views 2MB Size
THE JOURNAL OF

Organic Chemistry @ Copyright 1981 by the American Chemical Society

VOLUME46,NUMBER18

AUGUST28,1981

Synthesis of Iron Chelators. Enterobactin, Enantioenterobactin, and a Chiral Analogue William H. Rastetter,*’ Thomas J. Erickson, and Michael C. Venuti Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received March 31, 1981

Enterobactin (3), the iron-binding ionophore of the enteric bacteria, has been synthesized from L-serine. The antipode of the natural product, enantioenterobactin (42), has been analogously assembled from D-serine, and the iron-bindingand biological properties of the ligand have been investigated. Linear, chiral analogue 66 has been synthesized from L-asparagine,and the analogue Fe(II1) complex 67 has been characterized. Attempts to assemble the trilactam analogue 4 have failed at the cyclization step. T o contend with the insolubility of ferric hydroxide in microbes have an aqueous milieu [K,,(Fe(OH)3) = evolved specialized ligands or siderophores3to acquire and transport environmental iron into the cell. The siderophores bind Fe(II1) by wrapping six ligating atoms around the metal center in either a right-handed (A) or left-handed (A) coordination propeller (see enantiomeric complexes 1 and 2). We have assumed that the chirality of an iron3-

0

co

H

I , A-cis

Fe(lII).enterobactin

oc Qo

3-

H

2 , A-cis Fe(lll)* enantioenterobactin

siderophore complex may play a key role in the binding and transport of the complex into the microbial cell. To address the stereospecificity of the iron-gathering process in the enteric bacteria, we have synthesized the catechol-based siderophore enterobactin (3) and its antipode, enantioenterobactin (42, Scheme IV),4 from G and Dserine, respectively. Herein we report the details of these syntheses, including studies of racemization during the (1) Firmenich Assistant Professor of Natural Products Chemistry, Alfred P. Sloan Fellow, 1980-1982. (2) Linke, W. F. “Solubilities. Inorganic and Metal-Organic Compounds”, 4th ed.; Van Nostrand Princeton, NJ, 1958; Vol. 1, p 1039. (3) Raymond, K. N.; Carrano, C. J. Acc. Chem. Res. 1979.12.183 and . . reference; therein. (4) h t e t t e r , W. H.; Erickson, T. J.: Venuti, M. C. J. Om. Chem. 1980.

45, 5011.

0022-3263/81/1946-3579$01.25/0

3

Hb

4

HO

coupling of serine derivatives. The biological activity of enantioenterobactin is discussed in light of the chirality of its iron complex. Also discussed are strategies and difficulties encountered in an attempted synthesis of the trilactam analogue of enterobactin, 4, which is of interest in the study of the intracellular release of iron from iron(III)-enterobactin. Intermediates from this synthesis have yielded the linear, chiral analogue 66 (Scheme VII); its synthesis and the characterization of the complex Fe(II1)-66 (see 67) are reported. Background. Enterobactin5 is overproduced and excreted by Escherichia coli and related enteric bacteria under low iron stress. Solubilization of Fe(II1) is achieved upon formation of the exceptionally stable octahedral complex 1 (Kf = 1052).6 An outer membrane receptor protein then functions in the binding and transport of the siderophore-iron complex into the bacterial cell.’ In principle, two diastereomeric iron(II1)-enterobactin complexes can exist, the right-handed or A-cis complex (1) and the left-handed or A-cis complex (not shown).8 The in(5) (a) Pollack, J. R.; Neilands, J. B. Biochem. Biophys. Res. Commun. 1970,38,989. (b) OBrien, I. G.; Gibson, F. Biochim. Biophys. Acta 1970, 215, 393.

(6)Harris, W. R.; Carrano, C. J.; Cooper, S. R.; Sofen, S. R.; Avdeef, A. E.; McArdle,J. V.; Raymond, K. N. J . Am. Chem. SOC.1979,101,6097. (7) Hollifield, W. C., Jr.; Neilands, J. B. Biochemistry 1978,17,1922.

0 1981 American Chemical Society

3580

J. Org. Chem., Vol. 46, No. 18, 1981

teraction of the outer membrane receptor protein with each diastereomer might be quite different. Were both A& and A-cis enterobactin complexes available and separable, the sensitivity of membrane binding and transport to changes in metal center configuration could be evaluated. However, on the basis of the chromatographic behavior of kinetically inert enterobactin metal complexes and from correlations of CD spectra, Raymond and co-workersghave concluded that ferric enterobactin exists in solution as the A-cis isomer to the exclusion of the diastereomeric A-cis complex. By analogy, ferric enantioenterobactin must exist as the A-cis isomer. Membrane receptowiderophore complex interactions based on size, charge, and polarity must remain unchanged for iron(III)-m"noenterobactin, Bioassay of synthetic enantioenterobactin, thus, will reveal the dependence of binding and transport on solely the combination of platform and metal center chirality (natural, L-seryl and A-cis, 1, vs. unnatural, D-Seryl and A-cis, 2). Enterobactin is not reexcreted after the transport and iron-release proce~ses;~ i.e., a single ferric ion is transported by a molecule of enterobactin before the siderophore is hydrolyzed by the enzyme, enterobactin esterase (enterochelin esterase).1° The hydrolysis step may be a prerequisite for iron release.sJ1 Alternatively, it has been suggested12that the hydrolysis of the ester bonds is only coincidently related to the iron-delivery step and that iron release from iron(II1)-enterobactin is initiated by protonation and reduction of the complex to a neutral Fe(I1) species.13 Enterobactin analogue 4, lacking the hydrolytically labile14 seryl ester linkages of the natural product, might display some resistance to intracellular degradation by enterobactin esterase. Mutant E. coli cells deficient in the iron-release mechanism display a distinct pink coloration due to accumulated, highly colored iron(II1)enterobactin.lob Parent strain or mutant E. coli defective in the transport of iron(II1)-enterobactin are white and do not accumulate the red complex. It is intriguing to wonder whether E. coli with an intact iron-release system, but unable to rapidly hydrolyze the amide bonds of Fe(111)-4,would also accumulate the complex and develop a pink, ferric catecholate coloration. Thus, bacterial studies with trilactam 4 might reveal information relating the enzymatic dismembering of the enterobactin backbone to the iron-delivery step. Enantiomeric Enterobactins. Initial Approaches. The symmetry of enterobactin (3) dictates the synthetic (8)Were it possible to interchange the positions of the catecholate oxygens of one bidentate 2,3-dihydroxybenzoyl ligand in the A-cis complex 1, the A-trans complex would be formed; the A-cis and A-trans complexes are related analogously. Dreiding molecular models clearly show, however, that the trans complexes are too strained to form. (9) Raymond, K. N.; Kamal, A.-D.; Sofen, S. R. ACS Symp. Ser., 1980, NO.119, 133-167. VS. (10) (a) Greenwood. K. T.: Luke. R. K. J. Biochim. B ~ O D ~Acta 1978,525,209. (b) L&gman,'L.; Young, I. G., Frost, G. E., Risenberg, H.; Gibson, F. J. Bacteriol. 1972, 112, 1142. (11) Cooper, S. R.; McArdle, J. V.; Raymond, K. N. R o c . Natl. Acad. 1978., 75. 3551. Sci. ~ . U.S.A. . . --, ---(12) The carbocyclic analogue of enterobactin, cis-1,5,8-tris(2,3-dihydroxybenzamido)cyclododecane(Corey, E. J.; Hurt, S. D. Tetrahedron Lett. 1977,3923),is remarkably effective in iron deliver delivery E. coli. On the basis of this observation,HoWield and Neilands7have questioned the role of an esterase in.the metabolic event(s) involved in the release of the coordinated ferric ion from iron(III)-enterobactin. The formation constant for the Fe(II1) analogue complex, however, is not known, and, hence, the stability of the complex has not been directly compared to that of iron(III)-enterobadin (Kf= 1062).6 (13) Hider, R. C.; Silver, J.; Neilands, J. R.;Morrison, I. E. G.; Rees, L. V. C. FEBS Lett. 1979, 102, 325. (14) In aqueous buffer (TrieHCl, 0.5 M, pH 8.0) enterobactin is reported to be hydrolyzed substantially after 3 h at 37 "C; after 20 h the predominant hydrolysis product is monomer 10 (Scheme I): O'Brien, I. G.; Cox. G. B.; Gibson, F. Biochim. Biophys. Acta 1970,201, 453.

Rastetter, Erickson, and Venuti Scheme I Bn=benzyl

sw

-

EnOvCooH

1

7 H%En

:/c

eW*

HOXCooH

I . Zn, HOAc

2 H,,

Pd/C

I1

13,

D, L-dimer

0

14,

L,L-dimer

15,

D,L-dimer

approach to the siderophore;viz.,an efficient synthesis will entail the coupling and subsequent cyclization of protected L-serine monomer units. Analogously, Dserine might serve as chiral precursor to enantioenterobactin (42). Within these broad outlines we have evaluated two fundamentally different approaches to enterobactin and its antipode. The approaches differ in the timing of the catechol ligand attachment to serine nitrogen. The first entails the coupling of differentially protected N-[2,3-bis(benzyloxy)benzoyllserine monomers (for example 7 and 10, Scheme I), an approach which obviates the need for multiple deprotections and acylations of serine nitrogen late in the synthetic sequence. The second utilizes urethane protection for the monomer amino groups (e.g., see Scheme IV) and introduces the N-benzoyl ligands only after cyclization of the seryl ester platform. The potential problems and the ultimate outcome of each approach are discussed below. The risk inherent to any N-benzoyl monomer approach is the racemization of the activated ester required for monomer coupling. Nonetheless, the success achieved with dicyclohexylcarbodiimide/N-hydroxybenzotriazole (DCCLHOBt)l5 for the coupling of N-benzoyl amino acids' suggested the use of these reagents for the coupling of protected N-(2,3-dihydroxybenzoyl)serine monomers. Scheme I shows the synthesis of the monomers and the results of the coupling experiments performed in the Lserine series. Acylation of 0-benzyl-L-serine (5) with 2,3-bis(benzyloxy)benzoyl chloride (6)l6directly provides the alcohol(15) (a) Kbig, W.; Geiger,R. Chem. Ber. 1970,102,788. (b) Klaumer, Y.; Chorev, M. J. Chem. Soc., Chem. Commun. 1975, 973. (c) Chorev, M.; Knobler, Y.; Klausner, Y. J. Chem. Res., Synop. 1977, 202. (d) Weber, U. 2.Naturforsch., B: Anorg. Chem., Org. Chem. 1976,31,1157. (e) Young, G. T. "Proceedings of the 12th European Peptide Symposium"; Hanson, H., Jakubke, H., Eds.; North-Holland Publishing Co.: Amsterdam, 1973; p 132.

J. Org. Chem., Vol. 46, No. 18, 1981 3581

Synthesis of Iron Chelators

Scheme I11

Scheme I1

M;w;Br

HOTCOOH BocNH 16

17

+

HO/yCOOH

pB&COCH,Br+

BocNH 16

KHCO,

HTCOOCH2C00p&

Boc H 26

17 28

0, Pd/C

2

n P

BocNH BocNH

/

T H P O Y COOR

2 Zn,HOAc

BOCNH

2 7 , R = CH,COlpBr 28.

I DCC, H0Bt

BnOYCooH BocNH

I @,H*

___)

+

HVCooMaq BocNH 17

I DCC HOB1

2

H*

20, R=H

BnOd%$

R=H

ROgOY~ I 10, DCC, HOBt

BOC H BocNH 29, R = T H P 30, R = H

2

HP, Pd/C

YHBoc

N H N Bo’c ‘H Bok ‘Ht:N(C

J

6 II

2‘ 31, R , = H , R2=H

/HOBt

18

+

$b‘&)

BocNH

25

24

-

blocked monomer 7. Acid-blocked monomer 10 is made from enamine 817by alkylation and deprotection (8 9) followed by acylation (9 10). Both monomers are readily deprotected to give the enterobactin monomer l l.14J8 Amino acid analysislg of serine derived from 11 by hydrolysis shows both protected monomers (7and 10)to be of high chiral purity (>95% L). The coupling of monomers 7 and 10 with DCC/HOBt produces a 1:l mixture of diastereomeric dimers 12 and 13,separable by LC (silica gel). Fully deprotected dimers 14 and 15 show the chromatographic behavior (cellulose)of naturally derived dimer 14.14 Racemization,z0 usually held in check by the DCC/ HOBt procedure, becomes a rapid and undesired process with monomer acid 7. Neither changes in solvent nor reaction conditions give any improvement in the 1:l ratio of diastereomers formed in the coupling reaction. Apparently, the electron-releasing, o-benzyloxy group in monomer 7 promotes the nucleophilicity of the benzamide carbonyl oxygen and enhances the rate of azlactonization of the active ester formed from 7. Other coupling methods lead either to azlactone formation from 721aor to dehydration of monomer alcohol 10.zlM In view of these early difficulties with the N-benzoyl monomer approach, it was abandoned in favor of a urethane-protection route.

-

(16) Venuti, M. C.; Rastetter, W. H.; Neilands, J. B. J. Med. Chem. 1979, 22,123. 117) . la) . , Balon. A,: Breazu. D.: Daicoviciu. C.: Varea. E.: Beu. L.: Gonczy, F. Reu.%oum. Chim.‘1971,16,1601. ‘(b) MaclGen, J. A. Aust: J. Chem. 1972,25, 1293. (18) OBrien, I. G.; Cox, G. B.; Gibson, F. Biochim. Biophys. Acta 1969, 177,321. (19) Diastereomeric dipeptide analysis performed according to: Manning, J. M.; Moore, S. J.Biol. Chem. 1968,243,5591. We thank Drs. Ralph Hirschmann and Carl Bennett of Merck Sharp and Dohme Research Laboratories, West Point, PA, for carrying out these analyses. (20) Review: Kemp, D. S.In “The Peptides”; Grosa, E., Meienhofer, J., Eds.; Academic Press: New York, 1979; Chapter 7. (21) (a) DCC, pyridine. (b) N-Methyl-2-chloropyridiniumiodide, triethylamine: Mukaiyama, T.; Usui, M.; Shimada, E.; Saigo, K. Chem. Lett. 1975, 1045. (c) Ethyl diazodicarboxylate, triphenylphosphine: Mitsunobu, 0.; Eguchi, M. Bull. Chem. SOC.Jpn. 1971,44, 3427. (d) NJV’Jhrbonyldiimidazole: Staab, H. A.; Mannschreck, A. Chem. Ber. 1962,95, 1284.

32

Urethane protection for serine nitrogen avoids the problem of racemizationzoduring seryl ester formation.zz Inherent to this approach is urethane deprotection and introduction of N-benzoyl ligands after the complete assembly of the seryl ester platform. The required Nacylation of the three amino groups in the cyclic triester of L- or D-serine (see 40, Scheme IV), however, poses a potential risk late in the synthetic sequence owing to the propensity for 0- to N-acyl shifts in 0-acylserine derivat i v e ~ . This ~ ~ risk could be evaluated only through synthesis of the planned intermediate, cyclic 0-acylserine derivative 40. Schemes I1 and I11 show initial, yet ultimately unsuccessful, urethane approaches to enterobactin. As shown in Scheme 11, monomer alcohol 17 is formed (94%) upon with alkylation of N-(tert-butyloxy)carbonyl-~-serine~~ 2-(bromomethyl)anthraquinoneZ6(MaqBr). Coupling of 17 with 0-benzyl-N- [(tert-butyloxy)carbonyl]- serine (18)z7by the DCC/HOBt procedure produces crystalline dimer 19 in 80% yieldvZzCatalytic transfer hydrogenolysisa selectively and quantitatively removes the Maq ester, producing acid dimer 20. Despite the successful coupling of monomers, the coupling of acid dimer 20 with monomer alcohol 17 is not straightforward. The coupling produces a low yield of crystalline trimer 25 (19%), as well as crystalline dehydro dimer 24 (7%) and fully protected crystalline dimer 19 (35%). As shown in Scheme 11,these three products must arise via active ester 21 which can either couple (21 23 25)or fragment (21 18 22). The unexpected dimeric products 19 and 24 arise by coupling of the fragments (18and 22) with monomer alcohol 17. For avoidance of dimer fragmentation, serine oligomerization must be achieved by deprotection and chain extension at the seryl hydroxyl group (cf. carboxyl de-

--

-

+

I

(22) Dimer 19 (Scheme 11) was compared (TLC, ‘H NMR) with a diastereomeric dimer produced from racemic monomer acid 16 and optically pure L monomer 17. TLC and ‘H NMR showed no detectable diastereomeric dimer in the product from coupling of L-serine derived 16 and 17. (23) Schr6der, E.; Lfibke, K. “The Peptides”; Academic Press: New York, 1965; Vol. 1, pp 213-214. (24) The viability of the urethane approach was fvst shown by Corey in his enterobactin synthesis: Corey, E. J.; Bhattacharyya, S. Tetrahedron Lett. 1977,3919. (25) “Methoden der Organischen Chemie (Houben-Weyl)”;MUer, E., Ed.; Georg Thieme Verlag: Stuttgart, 1974; Band Teil I, p 132. (26) Kemp, D.S.;Reczek, J. Tetrahedron Lett. 1977, 1031. (27) Wuench, E.;Zwick, A. Chem. Ber. 1964,97, 2497. (28) Baer, E.;Maurukas, J. J. Biol. Chem. 1955,212, 25.

3582 J. Org. Chem., Vol. 46, No. 18, 1981

Rastetter, Erickson, and Venuti

Scheme IV THP.

1

DBU 33

2 hu,

I DCC,HOBt 34 + 3 5

'f OH

34

0 35

-

I 35, DCC, HOB1

HO+GOMaq

*

2 H+, EtOH

2 H*, EIOH 36

1 cyclization

H

O

0~

O

~

O 0 T

O

37, R=Maq 38, R = H

2 R H2, Pd/C HCI

i

L-4 I

X 3 9 , X=NHZ

I

40,

X=NH:

CI-

600 nm

S O 0 nm

400 nm

700 nm

Figure 1. Circular dichroism spectra: A,iron(III)-enterobadi; B,ferric complex 67;C,iron(II1)-enantioenterobactin. Scheme V

2 H,,

4 I , A r = 2,3-b~(benzylaxy)phenyl

Pd/C NHCOAr

A$-

Tos =

4 2 , A r = 2,3-dihydroxyphenyl

protection and extension, Scheme 11). Thus, monomer 28, made as shown in Scheme I11 (overall 76%), can be coupled with alcohol monomer 17, cleanly producing dimer 29. Deprotection yields crystalline alcohol dimer 30 (overall 80% from 17). Coupling of 30 with O-benzylblocked monomer 18 (from Scheme 11) gives the previously obtained (Scheme 11)protected trimer 25 as a crystalline solid (66%). Hydrogenolysis achieves deprotection of both ends of trimer 25, yielding hydroxy acid 31 (100%). U1timately, (tert-buty1oxy)carbonyl-protectedtrimer hydroxy acid 31 failed to cyclize cleanly under a variety of conditions, including those used by Corey et al.% for cyclization of the corresponding N-[ (benzyloxy)carbonyl]-protected trimer hydroxy acid (see enantiomer 38, Scheme IV). Crude cyclization mixtures derived from trimer 31 could not be purified nor deprotected to give a cyclic triammonium salt (see enantiomer 40, Scheme IV). Syntheses of the Enterobactins. Our successful approach to the enterobactins is outlined in Scheme IV for the D-seryl series. N-[(Benzyloxy)carbonyl]-~-serine~ (33) is alkylated with 2-(bromomethyl)anthraq~inone~~ (MaqBr), giving acid-protected monomer 34 (92%). Hydroxyl-protected monomer 35 is derived from 34 by reaction with dihydropyran followed by photoreductive deprotection% of the Maq ester (82% overall). Coupling of 34 and 35 and removal of the THP protective group yields crystalline dimer alcohol 36 (88% overall). Further coupling of 36 with 35 and T H P removal yields crystalline trimer alcohol 37 (95% overall). Removal of the Maq ester generates the enantiomer (38) of Corey's linear trimer enterobactin precursor24(6742% ). Trimer 38 can be cyclized in low yield by DCC/HOBt coupling, by Masamune's tert-butyl thioester/cuprous triflate method29 or, preferably, by Corey's imidazolyl disulfide procedure." Hydrogenolysis of crystalline 39 in the presence of HCl proceeds smoothly to trihydrochloride salt 40" which is acylated without isolation to give hexabenzylenantioenterobactin (41, 61-69% overall). The successful triacylation 40 41 and the similar conversion previously reported by Corey et al.24dispelled our fears concerning 0- to N-acyl shifts in the triamine liberated from salt 40. Hydrogenolysis of 41 yields enantioentero-

-

(29) Masamune, S.; Hayase, Y.; Schilling, W.; Chan, W. K.; Bates, G.

S. J . Am. Chem. SOC.1977,99, 6756.

Boc =

P

-+awe

benzyl

Bn

0 I No, NH,

43

44

pBr0CH2Br

HNm

B O ~ H N ~ ' ~ O H~~ ~ H3fTcoocH20ph u THF H@

OBn O B n

OBn OBn

47

4 5 , R=H 46, R=CH20pBr 45

+

47

DCC, HOBt

BocHN

Me

O h OBn

40, 49,

OBn OBn L,L-dimer D,L-dimer

bactin (41, 70-90%) which is indistinguishable from natural enterobactin by 'H NMR, IR, TLC, UV, field-desorption mass spectrometry (M', m / e 669), and melting point. Ferric enantioenterobactin (2) displays a CD spectrum which is the mirror image of that measuredm for the natural ferric complex (1, Figure 1). Scheme IV applied to N-[(benzyloxy)carbonyl]-~-serine~ yields enterobactin indistinguishable in all respects (vide supra) from natural material. Amide Analogues. Approaches to Trilactam 4. In place of serine a chiral source of a,@-diaminopropionicacid is required to adapt our enterobactin approaches to the assembly of analogue 4. L-Asparagine (43, Scheme V) fills this role and provides a convenient way of differentiating the a- and @-aminogroups of monomeric intermediates. Thus, literature procedures31convert asparagine (43) into (30) (a) Salama, S.; Strong, J. D.; Neilands, J. B.; Spiro, T. G. Biochemistry 1978,17,3781. (b) Rogers, H. J.;Synge, C.;Kimbar, B.; Bayley, P. M. Biochim. Biophys. Acta 1977,497, 548. (31) (a) Reader, C. J. Am. Chem. SOC.1960,82,1641. (b) Yoneda, N.; Fujii, T.; Uneda, M.; Yasuo, H.; Taguchi, Y.; Okurnura, K. Yakugaku Zasshi 1969, 89,98; Chem. Abstr. 1969, 70, 88228. (c) Pozdnev, V. F. Chem. Natl. Compd. (Engl. Transl.) 1974, IO, 782. (d) Hallett, A.; Wuensch, E.; Keller, 0.;Wersin, G. Hoppe-Seyler's 2. Physiol. Chem. 1976,357, 1651.

J. Org. Chem., Vol. 46, No. 18, 1981 3583

Synthesis of Iron Chelators Scheme VI

Known acid monomer 5032 is made from the Lasparagine-derived, chiral precursor 44. The amine monomer 52 is produced from 50 by alkylation (50 51, 87%) and p deprotection (51 52,87%). Coupling of monomers and amine deprotection affords crystalline dimer amine salt 54 (87% overall). Further coupling yields trimer 55; reductive cleavage of the Maq ester leads to crystalline trimer acid 56 (85% overall). Finally, cleavage of the terminal (tert-buty1oxy)carbonyl protective group provides the desired trimer amino acid 57 as a white, crystalline solid. To examine procedures which might be applied to the cyclization of trimer 57,we performed closure reactions with model trimer 59. Tri-0-alanine hydrochloride (59)%

I.Na NH L

COOH DBU BOCHN"~ ZNH Maq-Br

2 . OA0COCl

Tos H

44

~

50

52

51

O NHHN \ n NO

+,,&JN&OW53, X - NHBoc 54, x = NH: CF~CO;

50 ,DCC,HOBt

A

*

w

No&=

HCI

NaHC03

h&

~

EtOAc

/

TXZ

55, X=NHBoc, R = Maq 56, X = NHBoc, R = H 57, x = NH: CI-, R = H

ZHN

58

monomer 4432via tosylation, Hofmann rearrangement, and (tert-buty1oxy)carbonyl protection of the @-nitrogen. a deprotection of 44 by reductive cleavage of the tosyl sulfonamide allows the introduction of the N-[S,B-bis(benzyloxy)benzoyl]group or of urethane protective groups at the a-nitrogen. The former option allows a further test of racemization during the coupling of monomers bearing the protected 2,3-dihydroxybenzoylligand (cf. Scheme I). The syntheses and coupling of the N-benzoyl monomers is shown in Scheme V. Acid monomer 45 is produced (84%) upon a deprotection of 44 and acylation of the a-amine with acid chloride 6. From 45,monomer amine salt 47 is formed upon carboxylate alkylation (45 46,9870)and p deprotection (46 47,84%). The DCC/HOBt-mediated coupling of 45 and 47 affords a 1:l mixture of diastereomeric dimers 48 and 49 in quantitative yield. The amine-active ester coupling is slow relative to active ester racemization. Here, as in the serine acid monomer (7,Scheme I), the N-[bis(benzyloxy)benzoyl] group apparently promotes azlactonization and the lcas of monomer chiral integrity. The composition of the mixture 48/49 is verified upon comparison with authentic diastereomeric dimer made from racemic acid monomer (cf. 45) and monomer amine salt 47. The success of the (benzy1oxy)carbonylprotective group for serine nitrogen in the enterobactin synthesis (Scheme IV) prompted the synthesis of linear trimer amino acid 57 (Scheme VI). (Benzy1oxy)carbonyl-protected trimer 57, thus, is analogous to the enantioenterobactin linear precursor 38. Cyclization of 57 would afford the trilactam platform (see 58) required for analogue 4. The monomer syntheses and coupling steps for the urethane route are shown in Scheme VI.

-

-

-

Boc =

-

(32) Moore, S.; Patel, R. P.; Atherton, E.; Kondo, M.; Meienhofer, J. J. Med. Chem. 1976,19, 766.

"3N X-

H

59

H

CGY d

60

cyclizes in low yield to cyclo(tri-0-alanine) with DCC/HOBt (8%)or diphenylphosphoryl azide/triethylamine% (14%). Application of these procedures to trimer 57 produces only polymeric material, even at high dilution. Cyclizations with o-phenylene phosphorochloriditP or via the p-nitrophenyl ester37also fail to produce the desired cyclic material (58). Sterically, active esters from linear trimer 57 resemble active esters from enterobactin precursor 38 (Scheme IV). The failure to achieve the trilactam cyclization may reflect conformational rather than steric differences between the ester and amide precursors. It is interesting to note that a Dreiding or Corey-PaulingKoltun (CPK) model of trilactam 58 is easily constructed with all three amide linkages in the trans conformation. Thus, even the trans,trans conformation of linear diamide 57 should be able to cyclize. Nonetheless, intermolecular condensation (polymerization) proceeds faster than cyclization of trimer 57. Despite the inability to cyclize trimer 57,an analogue of enterobactin might be built upon the linear backbone of the trimer. Introduction of 2,3-dihydroxybenzoyl ligands at the a-nitrogens (see 66,Scheme VII) and chelation of Fe(II1) might yield a complex geometrically very similar to iron(III)-enterobactin (I).The iron center of the complex presumably would hold the ends of the linear backbone in close proximity, forming a nearly circular array (see 67). Further, the chirality of the backbone, as in enterobactin, might influence the chirality of the iron center. The transmission of chiral information from backbone to metal center, in principle, might be very different for the h e a r analogue with ita different flexibility and greater degrees of freedom. The synthesis of the h e a r analogue (66)is shown in Scheme VII. Trimer acid 56 (from Scheme VI) is converted into trimer ester 61 by the action of diazomethane; hydrogenolysis and acylation with 2,3-bis(benzyloxy)benzoyl chloride (6)affords fully protected trimer 63 (54% overall). Terminal amine deprotection (91%) and acetylation (72%) provides hexabenzyl-protected trimer 65. Hydrogenolysis of 65 yields the desired analogue 66 (75%). (33) Adams, E.; Davis, N. C.; Smith, E. L. J. B i d . Chem. 1952,199, 845. (34) (a) Rothe, M.Acta Chim. Acad. Sei. Hung. 1959, 18, 449. (b) Rothe, M.; Rothe, I.; Brung, H.; Schwenke, K. D. Angew. Chem. 1969,71, 700. (35) Brady, S. F.; Varga, S. L.; Freidinger, R. M.; Schwenk, D. A.; Mendlowski, M.; Holly,F. W.; Veber, D. R. J. Org. Chem. 1979,44,3101. (36) Rothe, M.; Kreiss, W. Angew. Chem., Int. Ed. 1973, 12, 1012. (37) Bohman-Lindren,G.; Ragnarseon,U. Tetrahedron 1972,28,4631.

3584

J. Org. Chem., Vol. 46, No. 18, 1981

Rastetter, Erickson, and Venuti

Scheme VI1 0

0

Boc =

CHfl,

56

$d B

QIAou

Z =

Ar:

01 o

c ZNH

H

N

q

ZNH 61

M

ZNH

e

H,

W/C=

THF, HOAc

L\rCOCI (6)-

Et3N, T H F

of a hydroxamate siderophore of fungal origin. The uptake of 55Fe(III)-enantioferrichromeis roughly 25-50% as efficient as the uptake of the natural siderophore complex. The transport study with enantioenterobactin (42) shows high chiral specificity of the outer membrane receptor protein for the natural siderophore complex (1) having an L-seryl platform and a A-cis metal center. T h e membrane chiral specificity for the natural complex may be determined by the chirality of the seryl triester backbone, by the chirality of the metal center, or by both of these factors. T h e large diameter of the chiral iron(II1)-catecholate center ( 14 A) and the insensitivity of the membrane transport process to changes in the platform16 suggest that the configuration of the metal center is the more important factor in determining membrane recognition. In further support of this, a synthetic, achiral, carbocyclic analogue of enterobactid2 forms a racemic Fe(II1) complex (50% A-cis and 50% A-cis) which is almost exactly half as effective as ferric enterobactin in competition for the solubilized outer membrane receptor of E. This observation is consistent with membrane receptor binding of only the A-cis complex, assuming slow equilibration of the two antipodes under conditions of the e ~ p e r i m e n t . ~ ~ N

62, 63,

65,

x = NH:

ACO-

X = NHCOAr

X = NHCOMe

Experimental Section

Fe(II1) complex 67 forms upon mixing 66 and FeC13 in aqueous methanolic buffer at pH 7.2. The shape, sign, and intensity of the CD curve for 67 are nearly identical with t h a t observed for iron(II1)-enterobactin (Figure l),indicating the A-cis geometry for the linear analogue complex (67). T h e formation constant for complex 67 has been measured at p H 7.2 by a competition experiment using ethylenediaminetetraacetic acid (EDTA) as an Fe(1II) scavenger. T h e absorbance of red complex 67 (Amm 495 nm) is 4%, 9%, and 8670,diminished a t 67/EDTA ratios of l:lO, 1:100, and 1:1000, respectively. From these data is calculated a formation constant for 67, Kf = Analogue complex 67 is, thus, some 5-6 orders of magnitude less stable than the natural complex 1 (K, = 1052).6 Iron-Transport Studies. Synthetic enterobactin (3), derived from L-serine, shows the full biological activity of naturally produced materia1.38*42Thus, synthetic 3 supports growth of mutant RW193 of E. coli K-12%which is deficient in the biosynthesis of the natural siderophore. Enantioenterobactin (42), in contrast, does not support bacterial growth; rather, 42 shows inhibitory activity toward strain RW193.42 Fe(III), tightly sequestered in the form of D-seryl A-cis complex 2, is unavailable to the bacterial transport system. T h e presence of siderophore antipode 42 in the growth medium thereby effectively prevents bacterial growth through iron starvation. This result is more dramatic t h a n t h a t observed recently in transport studies with enantioferrichrome,@the antipode (38) Details of the biological studies are discussed by: Neilands, J. B.; Erickson, T. J.; Rastetter, W. H. J. Biol. Chem. 1981,256, 3831. (39) American Type Culture Collection, No. 33475. (40) Winkelman, G. FEBS Lett. 1979, 97, 43. (41) The rate of racemization of the kinetically inert Fe(II1) complex (see ref 9) is not known. (42) We thank Professor J. B. Neilands (Berkeley) for testing of the enantiomeric enterobactins.

General Methods. All reactions were performed under a nitrogen atmosphere unless otherwise noted. Photolyses were carried out in Pyrex glassware by using a Rayonet-Srinivasan Griffin photochemical reactor (Southern New England Ultraviolet Co.) using RPR 3500-A lamps. Hydrogenations at 40-50 psi of H2 were performed in a Parr apparatus. Melting points are uncorrected and were determined in open-ended capillary tubes on a Mel-Temp or Thomas-Hoover melting point apparatus. 'H NMR spectra at 60 MHz were recorded on either a Varian T-60, Perkin-Elmer R-24B, or JEOL FX60 Q instrument; 90-MHz 'H NMR spectra were obtained on a JEOL FX90 Q and 250-MHz 'H NMR spectra on a Bruker WM-250 spectrometer. Chemical shifts are reported on the 6 scale relative to tetramethylsilane. Infrared spectra were recorded on either a Perkin-Elmer 567 or Perkin-Elmer 283B spectrometer. Circular dichroism (CD) spectra were recorded on a Cary 60 spectrometer, and optical rotations measured with a Perkin-Elmer 141 polarimeter. Ultraviolet and visible spectra were recorded on either a Perkin-Elmer 554 or Perkin-Elmer 330 spectrophotometer. Low-resolution mass spectra were determined on a Varian MAT 44 and high-resolutionmass spectra on a CEC llOB Mattauch-Herzog (Du Pont Instruments) mass spectrometer. Field-desorption mass spectra were determined on a Varian MAT 731 by using carbon emitters. Analytical TLC was carried out on E. Merck 0.2" F-254 silica gel on plastic backing, E. Merck 0.25-mm F-254 silica gel on glass plates, or Analtech 0.25-mm GF silica gel on glass plates. Compounds were detected by using W, iodine vapor, or 7% ethanolic phosphomolybdic acid char. Analytical chromatograms of enterobactin and other catechol-containing analogues were developed on Eastman cellulose on aluminum plates and visualized with 3% ferric chloride in 1 N HCl(aq). Preparative thin-layer chromatography was performed on E. Merck silica gel F-254 plates (20 cm X 20 cm X 2.5 mm). Gravity column chromatography was carried out on E. Merck silica gel 60 (7Cb230 mesh), and flash chromatography was performed according to Stilla on E. Merck silica gel 60 (230-400 mesh). Liquid chromatography (LC) was carried out on a Waters ALC/GPC 204 instrument using a p Porasil analytical column and UV detection (254 nm) or on a Waters Prep-BOO instrument using dual silica gel cartridges. Benzene, toluene, NJV-dimethylformamide, pyridine, acetonitrile, and CHzClzwere statically dried over 4-A molecular sieves. N-Methylmorpholine was distilled from BaO or dried over 4-A molecular sieves; Em was distilled from CaH2. THF' was distilled from Na/PhCOPh ketyl; DCC was distilled in vacuo. Anhydrous Et20 (Mallinckrodt) was used from sealed cans. 2,3-Di(43) Still, W. C.; Kohn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923.

Synthesis of Iron Chelators hydroxybenzaldehyde and 0-benzyl-L-serinewere obtained from Sigma Chemical Co. The following abbreviations are used: DMF = NJV-dimethylformamide, DCC = N~-dicyclohexylwbodiimide,DCU = NJV'-dicyclohexylurea, HOBt = N-hydroxybenzotriazole, 18crown-6 = 1,4,7,10,13,16-hexaoxacyclooctadecane, DBU = 1,8diazabicyclo[5.4.0]undec-7-ene, THF = tetrahydrofuran, MaqBr = 2-(bromomethyl)anthraquinone,Boc = (tert-butyIoxy)wbonyl, Me2S0 = dimethyl sulfoxide. 2,3-Bis(benzyloxy)benzoyl Chloride (6). To a solution of 2,3dihydroxybenzaldehyde(50.0g, 362 mmol) and benzyl chloride (100mL, 868 "01) in absolute EtOH (450 mL) was added KzCO3 (62.5g, 452 mmol). The mixture was refluxed 18 h; removal of solvent in vacuo and extractive workup (EtOAc/H20)provided crude 2,3-bis(benzyloxy)benzaldehydeas a moist solid TLC (silica gel, 3:l CH2C12-EtOAc) Rf 0.72. The crude aldehyde (ca. 362 mmol) was dissolved in acetone (350mL) and the solution diluted with HzO (350mL). To the cloudy mixture were added sulfamic acid (HZNS03H, 49.2 g, 507 mmol) and sodium chlorite (80%NaC102, 43.0 g, 380 mmol) in portions over 30 mimu The resulting solution was stirred 1 h at ambient temperature in an open vessel. Removal of acetone in vacuo yielded crude 2,3-bis(benzyloxy)benzoic acid as a precipitate in the remainingaqueous mixture. The solid was collected, washed with absolute EtOH, dried, and recrystallized from aqueous EtOH to yield pure 2,3-bis(benzyloxy)benzoic acid 113.4 g (94%); mp 124-124.5 "C (lit.'6 mp 124 "C); 'H NMR (60MHz, acetone-&) 10.5 (s, 1 H), 7.05-7.75 (m, 13 H), 5.28 (s,2H), 5.25 (s,2H); IR (KBr) 3030 (br), 1688,1572,1470,1450,1410,1375, 1300,1255,1215,1030cm-'; mass spectrum, m / e 344 (M+),243 (M' - C7H7), 225 (243- HZO), 91 (C7H7'). To a suspension of the acid (3.34g, 10.0 m o l ) in dry benzene (20 mL) at 5 OC was added DMF (3 drops) followed by oxalyl chloride (3.5mL, 40 mmol) dropwise. The resulting solution was stirred 30 min and then evaporated in vacuo to an oil. Generally the oil was used directly for acylations, but the acid chloride could be crystallized by trituration with 201 petroleum ether-Eb0, yielding 3.33 g (94%) of 6 mp 55-56 "C;'H NMR (60 MHz, CDCld 7.1-7.5 (m, 13 H), 5.1 (s,4H); IR (KBr) 1775,1685,1570, 1450,1375,1310,1265,1225cm-'; mass spectrum, m/e 352 (M'), 289 (M' - COssCl), 261 (M' - C7H7), 91 (C7H7+). 317 (M+ N-[2,3-Bis(benzyloxy)benzoyl]Monomer Acid 7. To a chilled solution of 0-benzyl-L-serine (5;10.0 g, 51.3 mmol) and NaOH (4.10g, 102.5 mmol) in H20 (100mL) was added dropwise a solution of acid chloride 6 (17.6g, 50.0 mmol) in THF (30mL). The mixture was stirred 1 h at 0 OC and 2 h at ambient temperature and then acidified to pH 2 with concentrated aqueous HCL Extractive workup (EtOAc/H20) and recrystallization from EtOAc yielded 20.0 g (78%) of monomer acid 7: mp 121 "C; 'H NMR (60MHz, CDClS) 9.25 (br 8, 1 H), 9.05 (d, 1 H), 6.90-7.85 (m, 18 H), 5.18 (8, 4 H), 5.00 (m,1 H), 4.42 (s, 2 H), 3.85 (m, 2 H);IR (KBr) 3360,3050(br), 1755,1635,1575,1538,1455,1265, cm-'; mass spectrum, m/e 511 1195,1140,1105,1052,1040,960 (M+), 420 (M' - C7H7), 402 (420- HZO), 358 (402- COZ), 91 (C7H7') Monomer acid 7 (1.10g, 2.15 mmol) was fully deprotected by hydrogenolysisover 10% Pd on carbon (110mg) in EtOH-HOAc (20:1,10mL). Filtration of catalyst, evaporation of solvent, and extractive workup [EtOAc/HC1(5% as), NaCl(aq) wash] yielded monomer 11 (0.480 g, 93%) as an off-whitepowder displaying ["ID +31.2O (c 2,MeOH). Hydrolysis of the monomer to give serine was achieved over 20 h a t 110 OC in 6 N HCl(aq). Coupling of the serine so produced with L-leucine-N-carboxyanhydrideand amino acid analysis of the resulting diastereomeric dipeptide^'^ showed the serine to be 4.1% D and 95.9% L. L-Serine subjected to the same hydrolysis conditions analyzed for 1.2% Dserine. On adjustment for serine racemization during hydrolysis, the chiral purity of monomer 11 (monomer 7)thus was 297.1% L [100% - (4.1% - 1.2%)]. N-[2,3-Bis(benzyloxy)benzoyl] Monomer Alcohol 10..A modification of the literature procedure17was used. LSerine ethyl acetoacetate enamine potassium salt 8 (mp 175-177 O C ; 20.3g, (44)Lindgren, B. 0.; Nilsson, T . Acta Chem. Scand. 1973, 27, 888. (45) Men, K. W . ;Fink, J. Arch. Pharm. (Weinheim, Cer.) 1956,289, 347.

J. Org. Chem., Vol. 46, No. 18, 1981 3585 79.5 mmol) was suspended in CH&N (750 mL), and to the mixture were added p-bromophenacylbromide (22.1g, 79.5 m o l ) and 18-crown-6(1.0g, 3.8 mmol). The suspension was refluxed 30 min, cooled to ambient temperature and filtered to remove precipitated salts. The filtrate was evaporated, and the solid residue was dissolved in freshly prepared HBr-MeOH (100mL, saturated) and stirred 10 min. The dark mixture was quickly evaporated to a solid residue and recrystallized from 2-propyl alcohol to yield hydrobromide salt 9 (29.0 g, 95%) as a tan solid mp 162-163 OC; 'H NMR (60MHz, D20) 7.43 (AB q, 4 H), 5.50 (8, 2 H), 4.72 (br s, HOD), 4.50 (t, 1 H), 4.12 (d, 2 H); IR (KBr) 3385,3220,2905,1740,1690,1580,1510,1430,1400,1365,1272, 1215,1105,1065,1045,1005,962cm-*. To a stirred solution of hydrobromide salt 9 (20.0g, 52 m o l ) and diisopropylethylamine (27mL, 155 mmol) in dry THF (200 mL) was added dropwise a solution of acid chloride 6 (17.9g, 50.9 mmol) in THF (ca.30 mL). The mixture was stirred at ambient temperature for 2 h and then concentrated in vacuo to onequarter volume. Extractive workup [EtOAc/5% HCl(aq), 5% NaHC03(aq) and NaCl(sat,aq) washes] and recrystallization from EtOAc yielded monomer alcohol 10 21.20 g (60%); mp 117-118 "C; 'H NMR (60MHz, CDC13) 9.00 (d, 1 H), 7.0-7.9 (m, 17 HI, 6.20 (br s, 1 H), 5.30 (AB q, 2 H), 5.13 (2s, 4 H),5.0 (m, 1 H), 4.05 (m, 2 H); IR (KBr) 3470 (br), 3360,2915,1752,1700,1655, 1585, 1572, 1520, 1450, 1372,1260, 1190, 1070,965 cm-'; mass spectrum, m / e 617 and 619 (M+, 79Br and 81Br),526 (M+ - Br), 420 (M+- C8H,$rO). Monomer alcohol 10 was fully deprotected by Zn/HOAc reduction followed by hydrogenolysis. Thus, alcohol 10 (1.30g, 2.10 mmol) in glacial acetic acid was stirred with zinc powder (5.0 g, 77 mmol) for 2 h. The zinc was removed by filtration and the filtrate evaporated in vacuo. Acid-base extractive purification of the crude product [Et20/5% NaHC03(aq); aqueous layer washed with EhO, acidified to pH 2 with concentrated HC1, and extracted with EtOAc; organic layer washed with 5% HCl(aq) and NaCl(sat, as)] and hydrogenolysis (conditions as for 7, vide supra) gave monomer 11 (0.350g, 70%) as an off-white powder displaying ["ID +30.5O (c 2,MeOH). Serine derived from 11 (10) was 5.6% D and 94.4% L. On adjustment for serine racemization during hydrolysis (vide supra), the chiral purity of monomer 11 (monomer 10) thus was 195.6% L [100% - (5.6% - 1.2%)]. Diastereomeric N-[2,3-Bis(benzyloxy)benmyl]Dimers 12 and 13. To a solution of acid monomer 7 (5.11g, 10.0 mmol), alcohol monomer 10 (5.56g, 9.0mmol), and HOBt (2.70g, 17.6 mmol) in dry CHSCN at 0 "C was added DCC (2.60g, 12.6 mmol). The mixture was stirred 1 h at 0 "C and 1 h at ambient temwas added and the mixture perature; pyridine (1.7mL, 21 "01) stirred 48 h. Precipitated DCU was removed by filtration, and the filtrate was evaporated in vacuo. Extractive workup [EtOAc/5% HCUaq), 5% NaHCOS(aq)and NaCl(sat, aq) washes] and column chromatography (375g silica gel, 1:9 EtOAcCHzC12) yielded the mixture of h e r s 12/13 (6.60g, 66%) as a white foam: 'H NMR (60MHz, CDCld 8.87 (d, 2 H),6.90-7.95 (m, 29 H),5.28 (br s, 2 H), 5.10 (br s, 8 H), 4.40-5.00 (br m, 4 H), 4.21 and 4.28 (2s of equal intensity, 2 H), 3.52 and 3.70 (2m of equal intensity, 2 H); IR (KBr) 3360, 3050, 1755, 1600,1575,1510,1450,1370, 1265,1195, 1070,965cm-'. The dimers could be separated by analytical TLC (silica gel, 1:9 EtOAeCH2Clz,R 0.26 and 0.23) or by preparative LC (silica gel, 1:14 EtOAc-CkzClZ). Diastereomerically pure dimer samples obtained by LC showed the following: 'H NMR (60MHz, CDC13) for Rf 0.26 dimer 4.21 (s, PhCHz0CH2),3.52 (m, PhCH20CH2);for Rf 0.23 dimer 4.28 (a, PhCH20CH2),3.70 (m, PhCH20CH2). The mixture of dimers 12/13 could be fully deprotected by sequential treatment with Zn/HOAc and then H2and Pd/C, (see conditions for deprotection of monomer 10). Deproteded dimers 14/15 show the same chromatographic mobility as naturally derived dimer" (cellulose TLC, 5% NH4+HCOO- in 0.5% HCOOH(aq),Rf for 14/150.76;monomer 11 used as R, standard). N-Boc Monomer Alcohol 17. To a suspension of MaqBr% (30.1g, 100 mmol) and N-Boc-L-serineZS(16;23.6 g, 115 mmol) in dry THF (350mL) at reflux was added DBU (14.9mL, 100 mmol) over 30 min. Heating at reflux was continued for an additional 30 min, during which time the mixture became homogeneous. cooling to ambient temperature led to crystallization of DBUSHBr; the salt was removed by filtration, and the filtrates

3586 J. Org. Chem., Vol. 46, No. 18, 1981 were evaporated to one-third volume. Extractive workup [EtOAc/l M aqueous citric acid, 5% NaHC03(aq) and NaCl(sat, aq) washes] and recrystallization from EtOAc afforded monomer 17 (39.75 g, 93.5%) as yellow crystals: mp 143-144 OC, 'H NMR (60 MHz, CDC13) 8.0-8.4 (m, 4 H), 7.55-7.90 (m, 3 H), 5.60 (d, 1H), 5.35 (s, 2 H), 4.50 (m, 1 H), 4.05 (m, 2 H), 2.87 (br t, 1 H), 1.40 (s,9 H); IR (KBr) 3500,3380,2980,1750,1710,1675,1595,1525, 1390, 1370, 1355, 1330, 1295, 1225, 1158, 1065,932 cm-'; mass MaqH+), 221 (CISH9O2+, Maq+), spectrum, m/e 222 (C15Hlo02+, 57 (C4H9+);[aID -15.25' (c 0.8, EtOH). Anal. Calcd for CBHBN06 C, 64.93; H, 5.45; N, 3.29. Found: C, 64.71; H, 5.49; N, 3.18. N-Boc monomer alcohol 17 produced as outlined above had spectral properties and an optical rotation indistinguishable from those of 17 produced from potassium salt 8 (Scheme I) by sequential reaction with MaqBr and di-tert-butyl dicarbonate. The procedure from N-Boc-L-serine,however, is more conveniently accomplished. N-Boc Dimer Acid 20. To a solution of acid monomer laz7 (4.87 g, 16.5 mmol), alcohol monomer 17 (6.38 g, 15.0 mmol), and HOBt (3.33 g, 21.8 mmol) in dry THF (60 mL) at 0 "C was added DCC (3.75 g, 18.2 mmol). The mixture was stirred 1 h at 0 "C and 1h at ambient temperature. N-Methylmorpholine (3.62 d, 33.0 mmol) was added and the mixture stirred 18 h at ambient temperature; TLC showed no remaining alcohol 17. The reaction mixture was doubled in volume with EtOAc and fitered to remove DCU. Extractive workup [EtOAc/l M aqueous citric acid, 5% NaHC03(aq)and NaCl(sat, aq) washes] and recrystallizationfrom 1:l EtOAc-hexanes afforded the dimeric Maq ester 19: 8.10 g (77%); mp 129-130 "C; 'H NMR (60 MHz, CDCld 8.00-8.40 (m, 4 H), 7.50-7.90 (m, 3 H), 7.30 (s, 5 H), 5.50 (br s, 2 H), 5.30 (s, 2 H), 4.204.80 (m, 4 H), 4.47 (s,2 H), 3.75 (d of AB q, 2 HI, 1.40 (e, 18 H); IR (KBr) 3470,3380,2980,1755,1742,1715,1705,1680, 1595,1505,1458,1398,1372,1340,1330,1290,1250,1210,1170, 1160,935 cm-'; mass spectrum, m/e 222 (C1&1002,MaqH+), 221

(ClSHBOP+, Maq+), 91 (C&+), 57 (C4H9+). To a solution of dimer Maq ester 19 (5.27 g, 7.50 "01) in 31 EtOH-DMF (40 mL) at 0 "C were added 10% Pd on carbon (1.75 g), N-methylmorpholine (1.1 mL, 10 mmol), and 1,4-cyclohexadiene (5 mL, 52 mmol). The bright green suspension of catalyst was stirred overnight at ambient temperature; TLC showed no remaining Maq ester 19. Filtration and extractive workup [EtOAc/l M aqueous citric acid, NaCl(aq) wash] yielded an amorphous solid. Flash chromatography (CH2C12to 5% HOAc in CH2C12gradient) afforded dimer acid 20 (3.30 g, 91%) as a crisp white foam: 'H NMR (60 MHz, CDC13) 10.45 (s, 1 H), 7.25 (5, 5 H), 5.50 (2 br overlapping d, 2 H), 4.20-4.80 (m, 4 H), 4.45 (8, 2 H), 3.65 (d of AB q, 2 H), 1.40 (s, 18 H). Dicyclohexylammonium salt: mp 141-142 "C; IR (KBr)3380,2970,1745,1720,1705,1640, 1580, 1500, 1390,1370,1340, 1250,1170,1060 cm-'. Coupling of N-Boc Dimer Acid 20 and N-Boc Monomer Alcohol 17. To a solution of dimer acid 20 (1.01 g, 2.10 mmol), monomer alcohol 17 (0.85 g, 2.00 mmol), and HOBt (0.424 g, 2.77 mmol) in THF (10 mL) at 0 OC was added DCC (0.476 g, 2.31 mmol). The mixture was stirred 1h at 0 "C and 1h at ambient temperature. N-Methylmorpholine (0.46 mL, 4.2 mmol) was added and the mixture stirred 18 h at ambient temperature; TLC showed only a trace of remaining alcohol 17. Workup as for dimer 19 (vide supra) afforded an off-white foam. Column chromatography (silica gel, 1:9 EtOAc-CH2Clz) of the mixture yielded three crystalline products, dehydrodimer 24 (78 mg, 7 % ) , protected dimer 19 (493 mg, 35%), and trimer 25 (340 mg, 19%) (in order of elution), Data for 24: mp 93-95 OC; 'H NMR (60MHz, CDC13)8.00-8.30 (m, 4 H), 7.50-7.85 (m, 3 H), 6.85 (br s, 1H), 6.00 (br s, 1 H), 5.40-5.60 (m, 2 H), 5.30 (8, 2 H), 4.70 (m, 1H), 4.55 (m, 2 H), 1.40 (8, 18 H); IR (KBr) 3380, 3430, 1750, 1720, 1675, 1595, 1510,1325,1295,1165,1065 cm-'; TLC (silica gel, 1:9 EtOAcCH&l& R,0.71. Data for dimer 1 9 identified by 'H NMR and TLC (silica gel, 10% EtOAc in CH2Cla)cospot with authentic dimer (vide supra), R 0.44. Data for trimer 25: mp 159-161 "C; 'H NMR (60 MHz, dDC1,) 8.00-8.40 (m, 4 H), 7.50-7.90 (m, 3 H), 7.25 (8, 5 H), 5.20-5.90 (m, 3 H), 5.30 (AB q, 2 H), 4.20-4.90 (m, 7 H), 4.55 (s,2 H), 3.50-4.00 (m, 2 H), 1.40 (s,27 H); IR (KBr) 3380,2990,1755,1745, 1720,1680,1600,1510,1375,1300,1255, 1170, 1065 cm-'; TLC (silica gel, 1:9 EtOAc-CH2C12) R, 0.40. Satisfactory elemental analysis for 25 was obtained on the couplmg product of dimer 30 and monomer 18 (vide infra).

Rastetter, Erickson, and Venuti N-Boc Monomer Acid 28. To a solution of N-Boc-L-serine= (16; 3.50 g, 20.0 mmol) and p-bromophenacyl bromide (5.56 g, 20.0 mmol) in acetone (125 mL) was added solid KHC03 (2.0 g, 20 mmol). The suspension was stirred overnight; TLC showed only traces of phenacyl bromide remaining. Water was added to effect solution, and the mixture was evaporated in vacuo to one-quarter volume. Extractive workup [EtOAc/H20, 5% NaHCO,(aq) and NaCl(sat, aq) washes] afforded a white solid; trituration with boiling EbO yielded N-Boc phenacyl ester 2 6 6.65 g (85%);mp 137-138 "C; 'H NMR (60MHz, CDCld 7.40-7.90 (AB q, 4 H), 5.42 (br s, 1H), 5.42 (AB q 2 H), 4.55 (m, 1H), 4.05 (d of AB q, 2 H), 3.90 (m, 1 H, exchangeable), 1.40 (8, 9 H); IR (KBr) 3460,3420,2980,1770,1700,1675,1592,1530,1425,1370, 1285, 1230, 1160, 1070,970 cm-'; mass spectrum, m / e 403 and 401 (M+, W r and *'Br), 372 and 370 (M+ - CHzOH), 316 and 314 (372 and 370 - C4H9),216 and 214 (p-BrPhCOCH,OH+), 57 (C4&+). Anal. Calcd for Cl$I&rN06: C, 47.78; H, 5.01; N, 3.48, Br, 19.86. Found C, 47.60; H, 5.03; N, 3.33; Br, 19.95. To a suspension of alcohol 26 (20.1 g, 50.0 mmol) and dihydropyran (19 mL, 200 mmol) in CH2C12(250 mL) was added The mixture pyridinium p-toluenesulfonate (1.20 g, 5.0 "01). rapidly became homogeneous; after 1h at ambient temperature, TLC showed no remaining alcohol 26. Extractive workup [CHzClZ/Hz0,NaCl(sat, aq) wash] afforded diastereomeric THP-protected monomer 27 as a syrup: 'H NMR (60 MHz, CDCld 7.50-8.00 (ABq, 4 H), 5.55 (d, 1H), 5.35 (br s, 2 H), 4.65 (br m, 2 H), 3.50-4.20 (m, 4 H), 1.00-2.00 (m, 15 H). Without purification, monomer 27 was dissolved in 1:19 HOAc-EbO (500 mL), and to the stirred solution was added zinc dust (102 g, 1.5 mol) portionwise. After 1.5 h, TLC showed no monomer 27 remaining. The suspension was filtered to remove zinc, and the filtrates were poured into HzO (200 mL). Solid NaHC03 was added in small portions until C02 evolution ceased. Acid-base extractivepurification [Eb0/5%NaHCOa(as); aqueous layer acidified with KHSO, to pH 4.5 and extracted with EhO; organic layer washed with NaCl(sat, aq)] afforded acid 28 as a syrup. Evaporation from toluene three times removed traces of HOAc in the product, affording diastereomericmonomer acid 28 (14.81 g, 102%) as a thick syrup containing residual toluene: 'H NMR (60 MHz, CDC13) 10.90 (8,1 H), 5.60 (2 overlapping d, 1 H), 4.20-4.70 (br m, 2 H), 3.40-4.20 (m, 4 H), 1.20-2.00 (m, 15 HI. N-Boc Dimer Alcohol 30. To a mechanically stirred solution of acid 28 (21.68 g, 75.0 mmol), alcohol 17 (21.25 g, 50.0 mmol), and HOBt (15.3 g, 100 m o l ) in THF (200 mL) at 0 "C was added DCC (17.0 g, 82.5 mmol). The mixture was stirred 1 h at 0 "C and 1h at ambient temperature. N-Methylmorpholine (16.0 mL, 150 mmol) was added and the mixture stirred 24 h; TLC showed no alcohol 17 remaining. The volume of the mixture was tripled with EtOAc and the mixture briefly heated (steam bath) to d m l v e some gelatinous material. Filtration of precipitated DCU, extractive workup [EtOAc/l M aqueous citric acid, NaCl(sat, aq), 5% NaHCO,(aq), and NaCl(sat, aq) washes], and evaporation of the solvent yielded a damp solid (dimer 29) used directly in the deprotection step. A small sample of 29 was dried in vacuo and characterized mp 122-124 "C; 'H NMR (60 MHz, CDCla 8.00-8.40 (m, 4 H), 7.55-7.90 (m, 3 H), 5.50 (br m, 2 H), 5.30 (a, 2 H), 3.20-4.90 (2 m, 9 H), 1.10-2.00 (m, 24 H); IR (KBr) 3480, 2940,1755,1740,1720,1710,1680,1595,1570,1430,1300,1160, 1060, 1030 cm-'. To the damp, crude dimer 29 (ca.50 mmol) in 5 1 EtOH-THF (600 mL) was added pyridinium p-toluenesulfonate (1.25 g, 5.0 mmol), and the resulting solution was stirred overnight at 55 "C. The mixture was evaporated in vacuo to one-quarter volume and partitioned between E g o (500mL) and NaCl(sat, aq) (200 mL). The EgO layer was washed with NaCl(sat, aq) (5 X 100 mL), dried, and evaporated to yield a yellow solid. Recrystallization from E t 0 afforded dimer 3 0 24.70 g (81% based on alcohol 17); mp 99-100 "C; 'H NMR (60 MHz, CDC13) 8.00-8.50 (m, 4 H), 7.60-8.00 (m, 3 H), 5.30-5.80 (br m, 2 H), 5.35 (8, 2 H), 4.20-4.90 (m, 4 H), 3.90 (m, 2 H), 2.60 (t,1H), 1.40 (s,18 H); IR (KBr) 3400 (br), 2980, 2940, 1750, 1715, 1680, 1595, 1510, 1300, 1160, 1060 cm-'. Anal. Calcd for C31HsN2011: C, 60.78; H, 5.92; N, 4.57. Found: C, 60.52; H, 6.05; N, 4.41. N-Boc Trimer Acid Alcohol 31. To a solution of dimer alcohol 30 (24.5 g, 40.0 mmol), 0-benzyl-N-(tert-butoxy-

Synthesis of Iron Chelators carbonyl)-L-serine (18; 17.70 g, 60.0 mmol), and HOBt (12.1 g, 80.0 "01) in dry THF (150 mL) at 0 "C was added DCC (13.60 g, 66.0 "01). The mixture was stirred 1 h at 0 "C and 1 h at ambient temperature. N-Methylmorpholine (13.2 mL, 120 m o l ) was added and the mixture stirred at ambient temperature for 24 h; TLC showed no remaining dimer 30. The mixture was doubled in volume with EtOAc and precipitated DCU removed by fitration. Elxtractive workup as for dimer 29 (vide supra) and recrystallization from Eta0 afforded trimer 25 (23.11 g, 65%) indkthgubhable (melting point, 'H NMR,IR,TLC) from trimer produced from the coupling of dimer acid 20 and monomer alcohol 17 (vide Scheme 11). Anal. Calcd for C&&015: C, 62.08; H, 6.23; N, 4.72. Found: C, 61.96; H, 6.37; N, 4.62. A solution of protected trimer 25 (13.34 g, 15.0 mmol) in 2:6:1 THF-HOAc-H20 (225 mL) was mixed with 10% Pd on carbon (3.50 g), and the suspension was shaken under H2 (50 psi) overnight. The catalyst was removed by filtration, and the fitrates were evaporated in vacuo to give a brown oil. Acid-base extractive purification [EtOAc/5% NaHC03(as); aqueous layer washed with QO, acidified to pH 4.5 with NaHSO,, and extracted with EtOAc; EtOAc layer washed with NaCl(sat,as)] and recrystallization from Et&hexanea afforded trimer 31 (8.68g, 100%)in two crops: mp 113-114 "C; 'H NMR (60 MHz, acetoned,J 7.75 (br s, 1 H), 6.30 (br m, 3 H), 4.55 (br m, 7 H), 3.95 (m, 2 H), 2.65 (br s, 1 H), 1.40 (e, 27 H); IR (KBr) 3350 (br), 2980,1740,1720,1700,1510,1450, 1385,1365,1340,1290,1240,1210,1055 cm-'. Anal. Calcd for C%)41N,&: C, 49.74; H, 7.13; N, 7.25. Found C, 49.50, H, 7.09, N, 6.97. N-[ (Benzyloxy)carbonyl] Monomer Alcohol 34. To a refluxing solution of N-[(benzyloxy)carbonyl]-~~~e~ (2.03g, 8.49 m o l ) and MaqBr (4.00 g, 13.3 m o l ) in THF (35 mL) was added in THF (5 mL) dropwise a solution of DBU (1.27 mL, 8.49 "01) over a period of 30 min. The solution was refluxed an additional 30 min and purified by extraction [CHClS/l N HCl(aq), 5% NaHC03(aq) and NaCl(sat, aq) washes]. Silica gel flash chromatography (CH2C12to 1:l EtOAeCH2C12 gradient) of the crude product afforded monomer 34 (3.60 g, 92%) as a pale yellow solid: mp 174-175 "C (recrystallized from EtOAc); 'H NMR (90 MHz, CDC13) 8.24 (m, 4 H), 7.74 (m, 3 H), 7.28 (s,5 H), 5.72 (d, 1 H), 5.32 (8, 2 H),5.10 (8, 2 H), 4.50 (m, 1 H), 4.02 (m, 2 H), 2.32 (t, 1 H, exchangeable); IR (KBr) 3520,3290,3060,2950,1725,1690, 1670,1590,1545,1345,1325,1290,1260,1215cm-'. Anal. Calcd for CzsHzlN07: C, 67.97; H, 4.61; N, 3.05. Found C, 67.84; H, 4.69; N, 2.96. N-[ (Benzyloxy)carbonyl] Monomer Acid 35. Monomer 35 was previously reported by Corey and Bhattacharyya;% our preparation of 35 via alcohol 34 follows, To a stirred suspension of alcohol 34 (4.71 g, 10.3 mmol) and pyridinium p-toluenesulfonate (262 mg, 1.04 mmol) in CH2Clz (50 mL) at ambient After temperature WBB added dihydropyran (3.7 mL, 40.5 "01). 1 h the mixture had become homogeneous. The yellow solution was stirred an additional 1 h, evaporated in vacuo to a volume of ca. 50 mL, and purified by extraction [EBO/H20, NaCl(sat, aq) wash]. Evaporation of the organic layer in vacuo afforded the THP ether of monomer alcohol 34 in quantitative yield (5.60 9). The yellow solid was used in the next step without further purification. A solution of the THP ether (2.10 g, 4.00 mmol) and Nmethylmorpholine (2 mL, 18.1 mmol) in 2-propanol (315 mL, freeze-thaw degassed) was irradiated at 350 nm for 4 h. Evaporation of the solvent in vacuo and extractive workup [EtOAc/NaHC03(sat, aq); aqueous layer washed with EtOAc, acidified to pH 3.5 with cold 10% aqueous HaO,, and extracted with CHtC12; organic layer washed with NaCl(sat, aq)] afforded diastereomeric monomer acid 35 (1.03 g, 82%) as a pale amber oil: 'H NMR (60 MHz, CDClJ 9.9 (8, 1 H), 7.25 (s,5 H), 5.9 (d, 1 HI, 5.1 (s,2 H), 4.3-4.8 (m, 2 H), 3.2-4.2 (m, 4 H), 1.1-1.9 (m, 6 H);IR (neat, NaCl) -2300 (br), 1715,1510,1450,1410,1340, 1240, 1200 cm-'. N-[ (Benzyloxy)carbonyl] Dimer Alcohol 36. To a stirred solution of acid monomer 35 (0.90 g, 2.79 mmol), alcohol monomer 34 (0.843 g, 1.83 mmol), and HOBt (0.551 mg,3.09 "01) in THF (8 mL) at 5 "C was added DCC (0.59 mL, 3.09 mmol) dropwise. The mixture was stirred at 5 OC for 1 h and at ambient temperature for 1 h, and then N-methylmorpholine (0.31 mL, 2.81 mmol) was added. After being stirred 20 h, the mixture was

J. Org. Chem., Vol. 46, No. 18, 1981 3587 diluted with EtOAc (20 mL) and HOAc (0.02 mL, 0.35 mmol) and stirred an additional 30 min. Filtration of precipitated DCU and extractive purification [EtOAc/l M aqueous citric acid, NaHC03(sat,aq) and NaCl(sat, aq) washes] yielded the dimer THP ether (1.40 g, 100%) as a pale yellow foam. The crude dimer THP ether (1.40 g, 1.83 "01) and pyridinium p-toluenesulfonate (50 mg, 0.199 mmol) were stirred at 65 "C in EtOH (20 mL) for 20 h. The solvent was evaporated to one-third volume, and the precipitated product was filtered from solution, yielding dimer alcohol 36 (1.09 g, €@YO) as a pale yellow solid mp 87-90 "C (recryetallied from EtOH); 'H NMR (60MHz, CDClJ 8.2 (m, 4 H), 7.7 (m, 3 H), 7.2 (8, 10 H), 6.2 (d, 1 H), 5.9 (d, 1 H), 5.25 (s, 2 H), 5.05 (s, 2 H), 5.0 (s,2 H), 4.2-4.9 (m, 4 H), 3.9 (br s , 2 H), 3.35 (8, 1 H, exchangeable); IR (CHCld 3420,3020,1740, 1715,1670,1590,1500,1450,1320,1290cm-'. Anal. Calcd for CnHsN2011: C, 65.29; H, 4.74; N, 4.12. Found C, 65.08;H, 4.68; N, 4.01. N-[ (Benzyloxy)carbonyl] Trimer Acid Alcohol 38. To a stirred solution of dimer alcohol 36 (8.03 g, 11.8 m o l ) , monomer in THF acid 35 (5.72 g, 17.7 mmol), and HOBt (3.62 g, 23.6 "01) (35 mL) at 5 "C was added DCC (3.72 mL, 19.5 "01) dropwise. The mixture was stirred at 5 "C for 1 h and at ambient temwas perature for 1 h, N-methylmorpholine (3.90 mL, 35.3 "01) added, and stirring was continued an additional 20 h The mixture was diluted with EtOAc (200 mL), and precipitated DCU was removed by fitration. Extractive workup [EtOAc/l M aqueous citric acid, 5% NaHC03(aq)and NaCl(sat, aq) washes] afforded the trimer THP ether (11.63 g, 100%) as a pale yellow foam. The crude trimer THP ether (14.69 g, 14.90 mmol) and pyridinium p-toluenesulfonate (0.375 g, 1.49 mmol) in 15:24 CHC13-EtOH (195 mL) were stirred at 55 "C for 23 h The solvent was evaporated in vacuo to one-fifth volume, and the precipitated trimer alcohol 37 (12.80 g, 95%) was collected by fitration: mp 131-133 "C (recrystallized from EtOH); 'H NMR (60 MHz, CDClJ 8.2 (m, 4 H), 7.7 (m, 3 H), 7.2 (8, 10 H), 7.15 (s,5 H), 5.8-6.3 (m, 3 HI, 5.25 (8, 2 H), 5.05 ( 8 , 4 H), 5.0 ( 8 , 2 HI, 4.0-4.9 (m, 7 H), 3.8 (m, 2 H), 3.1 (s br, 1 H, exchangeable);IR (CHClJ 3420, 3030,1745,1715,1670,1590,1505,1450,1380,1320,1290 cm-'. Anal. Calcd for C,J-IaN~O15:C, 63.93; H, 4.81; N, 4.66. Found C, 63.79; H, 4.95; N, 4.60. A degassed (N2 swept, 30 min) solution of trimer Maq ester 37 (3.75 g, 4.15 "01) and N-methylmorpholine(2.3 mL, 21 m o l ) in 23:38 CHC13-2-propanol(305 mL) was irradiated at 350 nm for 4.5 h. The yellow solution was evaporated in vacuo and purified by extraction [EtOAc/l M citric acid(aq), NaCl(sat, aq) wash; organic layer diluted with EhO and washed with 5% NaHC03(aq) (10 X 60 mL); combined aqueous layers at 5 "C acidified to pH 3.5 with 10% H2S04(aq)and extracted with EtOAc; organic layer washed with NaCl(sat, aq)], yielding acid alcohol 38 (1.90 g, 67%;yield of trimer in L-seryl series, 82%) as an off-white foam: 'H NMR (250 MHz, CDC13) 7.28 (m, 15 H), 6.11-6.30 (m, 3 H), 5.25-5.85 (br s , 2 H), 5.09 (m, 6 H), 4.35-5.04 (m, 7 H), 3.95 (half an AB q, J = 11 Hz, 1 H), 3.66 (half an AB q, J = 11 Hz, 1 H); IR (CHClS) 3420,3020,2950,1740,1715,1505, 1450,1400,1335,1290,1230,1210,1055 cm-';mass spedrum (field desorption), m / e 681 (M'). The enantiomer of trimer 38 was previously reported by Corey and Bhattacharyya.% N-[ (Benzyloxy)carbonyl] Cyclic Triester 39. The cyclization of trimer 38 follows the general procedure previously reported;% details of our cyclization follow. A solution of trimer 38 (0.707 g, 1.04 mmol), 4-tert-butyl-N-isopropylimidazolyl diand triphenylphosphine (0.543g, 2.07 sulfide (0.Wg, 1.15 "011, "01) in dry CHzClz(8.3 mL) was stirred at ambient temperature for 15 min. The mixture was diluted with dry toluene (31 mL) and drawn into a 50-mL syringe cooled with powdered COz(s). The cold solution was added via syringe pump over 1.5 h to toluene (260 mL) stirred at 67 "C (internal temperature). After the addition, the solution was stirred an additional 1 h at 67 OC and evaporated in vacuo to a volume of 30 mL. The mixture was stirred with CH31 (25 mL, passed through basic alumina) at ambient temperature for 10 min and then allowed to stand at 0 "C for 12 h Precipitated solids were removed by fitration (&lite), and the fdtrate was evaporated to a volume of 15 mL. Extractive purification [EtOAc/l M citric acid(aq), NaCl(sat, aq), 0.2 M aqueous citrate buffer, (pH 5.51, and NaCl(sat, aq) washes] and flash chromatography (silica gel, 7:93 acetone-CHzClz)afforded

3588 J. Org. Chem., Vol. 46, No. 18, 1981 cyclic triester 39 (72 mg, 10%; yield of cyclic triester in L-seryl series, 16%)as a white foam: mp 182-183 OC (recrystallizedfrom EtOAc-EtzO); 'H NMR (250 MHz, CDCl3) 7.32 (8, 15 H), 6.30 (m, 1 H), 5.84 (d, 1 H), 5.10 (s,6 H), 4.70 (m, 6 H), 4.37 (m, 3 H), 3.90 (m, 1H); IR (CHCld 3420,3330,3020,2980,2880,1755,1720, 1505,1450,1380,1330,1290,1210,1105,1060cm-'; maw spectrum (field desorption), m / e 664 (M+). Anal. Calcd for C33H93N3012: C, 59.73; H, 5.01; N, 6.33. Found: C, 59.71; H, 5.08;N, 6.14. Rexabenzylenantioenterobactin(41). A suspension of 10% Pd on carbon (63 mg) and cyclic triester 39 (41 mg, 0.062 "01) in THF (4.2 mL) containing 5% (v/v) DMF was flushed with H2 After the addition of 1.72 N HC1 in dry THF (0.330 mL, 0.561 mmol), the mixture was stirred at ambient temperature under an atmosphere of H2for 12 h, and then the solvent was evaporated in vacuo. The residue, containing catalyst and triammonium salt 40,was suspended in THF (3 mL), and solutions of 2,3-bis(benzy1oxy)benzoylchloride (6;135 mg, 0.371 mmol) in THF (0.3 mL) and EbN (0.052 mL, 0.374 mmol) in THF (0.3 mL) were added simultaneously via syringe over 10 min at 0 OC. The mixture was stirred at 0 OC for 15 min and at ambient temperature for 30 min. The hydrogenation catalyst was filtered from solution (Celite), and the filtrate was purified by extraction [EtOAc/l M citric acid(aq), 0.2 M aqueous citrate buffer (pH 5.5) and NaCl(sat, aq) washes] followed by flash chromatography (silica gel, 7:93 acetone-CHzC12), affording a mixture of hexabenzylenantioenterobadin (41)and 2,3-bis(benzyloxy)benzoicacid. This mixture was dissolved in CH2C12(5 mL),and ethereal diazomethenewas added until a yellow color persisted. The solution was stirred 5 min at ambient temperature and quenched with a drop of HOAc. Evaporation in vacuo and flash chromatography (silica gel, 7:93 acetone-CHzClz) afforded hexabenzylenantioenterobactin(46 mg, 61%; yield of hexabenzylenterobactin in the L-seryl series, 69% as a white foam: 'H NMR (250 MHz, CDCl,) 8.48 (d, 3 H), 7.65 (m,3 H), 7.08-7.45 (m, 36 H), 5.15 (half an AB q, J = 11Hz, 3 H), 5.11 (s, 6 H), 5.04 (half an AB q, J = 11 Hz, 3 H), 4.91 (m, 3 H), 4.09 (m, 6 H); IR (CHC13 3350,3020,1755,1655,1570,1505, 1450, 1370, 1340, 1310, 1260, 1210 cm-'. Anal. Calcd for C72HmN3015: C, 71.45; H, 5.25; N, 3.47. Found C, 71.61; H, 5.44; N, 3.36. Enantioenterobactin (42). A suspension of 10% Pd on carbon (19 mg) and hexabenzylenantioenterobactin (41;45 mg, 0.0372 mmol) in 2:l THF-HOAc (2.1 mL) was shaken under H2(50 psi) for 20 h. The catalyst was removed by centrifugation and washed with EtOAc (5 X 3 mL). The combined supernatants were washed with 1 N HCl(aq) (4 X 4 mL) and 0.2 M aqueous citrate buffer (pH 5.5,3 X 4 mL), dried over Na$04, and evaporated in vacuo. The residue was filtered through a short column of silicic acid (Mallinckrodt, 100-mesh; 4:l benzene-EtOAc) to afford enantioenterobactin (17.4 mg, 70%; yield of enterobactin in the L-seryl series, 90%) as a white powder: mp 200.5-202 *C (recrystalliied from EtOAc-hexanes) [lit6"mp (for natural enterobadin) 202-203 "C]; 'H NMR (250 MHz, acetone-da)8.40 (d, 3 H), 7.20 (m, 3 H), 7.00 (m, 3 H), 6.72 (m, 3 H), 5.09 (m, 3 H), 4.71 (d, 6 H), 2.7-3.1 (br s, exchangeable H's); IR (KBr) 3700-2500 (br), 1745,1635, 1580,1530,1455,1340,1260,1170,1070cm-'; mass spectrum (field desorption), m / e 669 (M'), 446 (seryl dimer+), 223 (seryl mo316 (e 9200). Anal. Calcd for nomer+); UV (EtOH) A,, C30HnN3Ol5: C, 53.82; H, 4.06; N, 6.28. Found C, 53.50; H, 4.29; N,6.27. N-[2,3-Bis(benzyloxy)benzoyl]Monomer Acid 45. To a stirred refluxing solution of tosyl sulfonamide 4432(3.60 g, 10.05 mmol) in anhydrous ammonia (200 mL) was added sodium (1.17 g, 51.04 mmol) in small pieces. The dark blue solution was refluxed for 1h, B i e h d Ag 50W-X4 (NH4+form) ion-exchange resin (14.7 g) was added, and the NH3 was allowed to evaporate overnight. The residue was dissolved in H20 (100 mL), and the resulting solution was filtered. The filtrate was washed with CHC1, (2 X 100 mL) and the aqueous layer evaporated in vacuo to afford a white solid. To the solid and NaHCO, (2.12 g, 25.26 mmol) in H 2 0 (16 mL) was added dropwise 2,3-bis(benzyloxy)benzoyl chloride (6;3.64 g, 10.32 "01) in THF (5 mL). The mixture was stirred at ambient temperature for 1 h, diluted with H2O (100 mL), and washed with EhO (2 X 50 mL). The aqueous layer was acidified to pH 1 with 5 N HCl(aq) and extracted with EtOAc. The organic layer was washed with NaCl(sat, as), dried, and evaporated to yield acid monomer 45 (4.37 g, 84%) as a white

Rastetter, Erickson, and Venuti foam; dicyclohexylammonium salt, mp 133-135 "C (recrystallized from EtOAc-hexanes). Data for 45: 'H NMR (60 MHz, CDCl,) 8.70 (d, 1H), 6.9-7.8 (8, 1H, m, 13 H), 5.10 (m, 4 H, m, 1H), 4.60 (m, 1H), 3.45 (m, 2 H), 1.35 (s,9 H); IR (CHC13 3450,3350,3010, 2980,1710,1645,1570,1510,1455,1390, 1370,1260, 1160 cm-'. N-[2,3-Bis(benzyloxy)bnzoyl]Monomer Amine Salt 47. To a stirred solution of acid monomer 45 (744 mg, 1.43 mmol) in 1 0 1 MeOH-H20 (30 mL) was added a 20% (w/v) solution of Cs2C03(aq)until the solution reached pH 8 (required 1.5 mL, 0.92 mmol). The solution was evaporated in vacuo and reevaporated twice from DMF (2 X 10 mL). The residue was dissolved in DMF (11mL), pbromobenzyl bromide (355 mg,1.42 "01) was added, and the mixture was stirred at ambient temperature for 15 h. The solvent was removed in vacuo, and the residue was triturated with HzO and then purified by extraction [benzene/NaHCO,(sat, aq), NaCl(sat, aq) wash], affording p-bromobenzyl ester 46 (966 mg, 98%) as a white foam: 'H NMR (60 MHz, CDC13) 8.6 (d, 1 H), 7.0-7.7 (m, 17 H), 5.1 (s, 4 H), 5.05 (s, 2 H), 4.7 (2 m, 2 H), 3.45 (t, 2 H), 1.38 (8, 9 H). p-Bromobenzyl ester 46 (2.86 g, 4.15 "01) was stirred in 4.0 N HCl in THF (90 mL) for 1.5 h. Excess HCl was removed by sweeping the solution with N2,and then the solvent was evaporated in vacuo. The residue was triturated with EhO and recrystallized from EtOH, yielding hydrochloride salt 47 (2.17 g, 84%) as white crystals: mp 163-166 OC; 'H NMR (60 MHz, CDC13) 8.6 (d, 1 H), 8.0-8.6 (br s, 3 H), 6.8-7.6 (m, 17 H), 4.95 (m, 6 H), 2.9-3.8 (m, 3 HI; IR (KBr) 3280,3080,3060,3030,2870, 1730,1650,1595,1570,1510,1450,1385,1370,1355,1330,1270, 1220 cm-'. Diastereomeric N-[2,3-Bis(benzyloxy)benzoyl]Dimers 48 and 49. To a stirred solution of monomer acid 45 (475 mg, 0.912 mmol), amine salt 47 (576 mg, 0.920 mmol), HOBt (251 mg, 1.857 in THF mmol), and N-methylmorpholine (0.101 mL, 0.920 "01) dropwise. (10 mL) at 0 OC was added DCC (0.181 mL, 0.949 "01) The suspension was stirred at 0 "C for 1 h and at at ambient temperature for 2 h. Precipitated DCU was filtered from the solution and the filtrate evaporated in vacuo. Extractive workup [EtOAc/l N HCl(aq),NaHC03(sat,aq) and NaCl(sat, aq) washes] afforded a white form. The remaining DCU was removed by dissolving the dimers in EtOAc and filtering the precipitate. Evaporation of the solvent yielded the dimers 48/49 (1.036 g, 100%) as a white foam: 'H NMR (60 MHz, CDC13)8.5 (m, 2 H), 6.9-7.7 (m, 31 H), 5.05 (s, 8 H),5.0 (s, 2 H), 4.75 (m, 1H), 4.2-4.6 (m, 2 H), 3.1-3.8 (m, 4 H), 1.35 (s,9 H); IR (KBr) 3330,3080,3060, 3030,2970,2910,1735,1705,1660,1640,1575,1520,1450,1430, 1375,1310,1260,1210, 1190 cm-'. TLC (silica gel, 1:l hexanesEtOAc) showed the dimers as two closely running spob, R,0.21 and 0.25; analytical LC (p-Porasil, 3:l CH2C12-EtOAc) showed two peaks of equal area. The dimer mixture was indistinguishable from authentic diastereomeric dimer made from racemic acid (cf. 45) and chirally pure monomer 47. Fractional recrystallization of the mixture 48/49 gave two crops (mp 153-159 and 65-70 OC, respectively). The higher melting crop showed a 9 1 enrichment in the slower eluting peak (LC); the lower melting crop showed a 9:l enrichment in the faster eluting peak (LC). The 'H NMR (60 MHz, CDCl,) of each crop was indistinguishablefrom the 'H NMR of the crude 1:l mixture of 48 and 49. Monomer Amine Salt 52. To a refluxing solution of monomer and MaqBr (3.66 g, 12.2 "01) in THF acid 50 (2.57 g, 7.60 "01) (45 mL) was added dropwise a solution of DBU (1.14 mL, 7.62 mmol) in THF (5 mL) over 30 min. The mixture was refluxed an additional 30 min, cooled to ambient temperature, and purified by extraction [CHC13/1 N HCl(aq), NaHC03(sat, aq) and NaCl(sat, aq) washes]. The product was purified further by silica gel flash chromatography (CH2C12to 3 1 CH2C12-EtOAcgradient), affording fully protected monomer 51 (3.70 g, 87%) as a pale yellow solid mp 176.5-177.5 O C (recrystallized from EtOAc); 'H NMR (60 MHz, CDCl,) 8.3 (m, 4 H), 7.8 (m, 3 H), 7.3 (s, 5 H), 6.1 (d, 1H), 5.4 (s, 2 H), 5.2 (m, 1H), 5.15 (s, 2 H), 4.6 (m, 1H), 3.6 (m, 2 H), 1.4 (s,9 H); IR (KBr) 3350, 2980, 2960, 1740, 1670, 1590,1520,1440,1420,1385,1360,1350,1290,1260,1035 cm-'. A solution of monomer 51 (3.60 g, 6.45 mmol) in 7:3 CF3COOH-H20 (30 mL) was stirred at room temperature for 3.5 h and then evaporated in vacuo. Reevaporation from toluene (three times) and trituration with EhO afforded monomer amine salt 52 (3.20 g, 87%) as a pale yellow powder: mp 169-170.5 OC

Synthesis of Iron Chelators

J,Org. Chem., Vol. 46, No. 18, 1981

3589

mmol) was added via syringe. The solution was stirred at -40 OC, and Et3N was added until the pH, measured with premoistened, narrow-range pH paper, was greater than 7.0. The mixture was allowed to stand at -25 OC for 2 days and at 0 OC for 4 days. The pH was maintained above 7.0 during this time by the addition of Et& every 12 h The mixture was diluted with HzO (10 mL), stirred with Bio-Rad RG 501-XS (20-50 mesh) mixed-bed resin for 24 h, filtered, evaporated in vacuo, and recrystallized from H20 (1 mL) to afford cyclic trimer 60 (17.1 mg, 14.3%) as colorless n d e s : mp 7360O C ; 'H NMR (SoMHz,DzO) 3.5 (t,6 H), 2.5 (t, 6 H); IR (KBr) 3300, 3100, 2990, 2940, 2880, 2850,2480,2440,1660,1565,1445,1430,1370,1345,1285,1245, 1200 cm-'; mass spectrum, m / e 213 (M+),185 (M+ - CO or M+ - CZH4). (B) DCC/HOBt Closure. To a solution of tri-@-alaninehydrochloride (59;489 mg, 1.83 mmol) at 0 "C were added HOBt (327 mg, 2.14 mmol), N-methylmorpholine (200 pL, 1.81 mmol), The mixture was allowed to warm and DCC (0.38mL, 1.99 "01). to ambient temperature and was stirred for 110 h. Precipitated polymer was filtered from solution and the filtrate evaporated in vacuo, affording a white powder. Recrystallization from HzO yielded cyclic trimer 60 (32 mg,8%) as colorleas needles: mp >360 "C; other data, vide supra. Fully Protected N-[ 2,3-Bis(benzyloxy)benzoyl] Trimer 63. To a solution of trimer acid 56 (0.38 g, 0.49 mmol) in 9 1 CHzCl2-MeOH (20 mL) was added ethereal diazomethane until a yellow color persisted. After the mixture was stirred 10 min, 3340,3080,3040,2960,1750,1720-1660,1610,1595,1535,1460, 1330,1300,1210,1055,935 cm-'. Anal. Calcd for C&&OllF3: excess CHzNzwas quenched with a few drops of HOAc and the C, 59.09; H, 4.45; N, 7.07; F, 7.19. Found C, 59.38; H, 4.49; N, mixture evaporated in vacuo. Silica gel column chromatography 7.06; F, 6.81. (1:lEtOAc-CHzCIJ yielded trimer methyl ester 61 (340 mg,88%) as a white solid mp 155-157 OC; 'H NMR (250 MHz, CDC13) Trimer Amino Acid Salt 57. To a solution of monomer acid 7.32 (m, 15 H), 7.13 (m, 1 H), 7.01 (m, 1 H), 6.75 (m, 1 H), 6.21 50 (1.65 g, 4.89 mmol) and HOBt (925 mg, 6.05 mmol) in DMF (m, 2 H), 5.12 (m, 6 H, m, 1 H), 4.47 (m, 1 H), 4.26 (m, 2 H), 3.98 (8 mL) at 0 "C was added DCC (0.96 mL, 5.03 mmol) dropwise. (m, 1 H), 3.78 (s,3 H, m, 2 H), 3.53 (m, 2 H), 3.30 (m, 1 H), 1.35 The mixture was stirred at 0 OC for 1 h and at ambient tem(s,9 H); IR (KBr) 3300,3080,3060,3030,2970,2950,1740-1640, perature for 1 h, and a solution of amine salt 54 (3.83 g, 4.83 "01) 1520,1450,1430, 1385,1360,1340, 1245, 1160, 1040 cm-'. and N-methylmorpholine (0.54 mL, 4.89 mmol) in DMF (8 mL) A suspension of trimer methyl ester 61 (200 mg, 0.252 mmol) was added. The mixture was stirred 20 h, warmed to dissolve precipitated trimer, and filtered to remove precipitated DCU. The and 10%Pd on carbon (140 mg) in 3 2 THF-HOAc (50 mL) was filtrate was evaporated in vacuo and the residue purified by flushed with Hz and shaken at 40 psi of Hz for 18 h. The catalyst extraction [CHC13/1M aqueous citric acid, NaHC03(sat,aq) and was filtered from solution, and the filtrate was evaporated in vacuo, NaCl(sat,aq) aq) washes], yielding trimer 55 (4.29 g, 89%) as a stripped from toluene (three times) in vacuo, and triturated with pale yellow solid. EtOAc to afford 130 mg of triammonium acetate 62 (91%)as a white powder: 'H NMR (60 MHz, DzO + CD,CN) 4.4 (exCrude trimer 55 (5.38g, 5.39 "01) in freshly distilled, degassed changeable H s and methine protons), 3.2-4.0 (m, 6 H, s, 3 H), 1,4dioxane (300 mL) was treated with a solution of Na#z04 (12.2 1.95 (8,9 H), 1.45 (s, 9 H). g, 70.1 "01) in 5% NaHCO, (150 mL, pH of solution 7.2). The mixture was stirred at ambient temperature for 1 h, HzO (100 To a stirred suspension of triammonium acetate 62 (190 mg, mL) was added, and stirring was continued an additional hour. 0.333 mmol) in THF (5 mL) at 0 O C were added simultaneously Dioxane was removed in vacuo, and the resulting aqueous mixture solutions of 2,3-bis(benzyloxy)benzoyl chloride (6;788 mg, 2.14 was extracted with EtOAc [ l M aqueous citric acid, HzO and mmol) in THF (2 mL) and EbN (0.290 mL, 2.09 mmol) in THF NaCl(sat, aq) washes]. The residue from evaporation of the (2 mL) over a period of 20 min. The resulting suspension was organic layer was purified by silica gel column chromatography stirred at 0 "C for 10 min and at room temperature for 1 h. (5% HOAc in CHzClzto 5% HOAc in 3 1 EtOAc-CHzCl2 graExtractive workup [EtOAc/l M aqueous citric acid, NaHC03(sat, dient), affording trimer acid 56 (4.03 g, 95%)as a white powder: aq) and NaCl(sat, aq) washes] and silica gel flash chromatography mp 158-160 OC (recrystallized from EtOH-EtzO); 'H NMR (60 yielded a mixture of trimer 63 and 2,3-dihydroxybemicacid. This MHz, CD30D + MezSO-d6)7.4 (s, 15 H), 5.2 (8,6 H), 4.55 (exmixture was dissolved in CHzClz(10 mL), treated with ethereal changeable Hs), 4.1-4.5 (m, 3 H), 3.3-3.7 (m, 6 H), 1.45 (s,9 H); diazomethane until a yellow color persisted for 10 min, quenched lR (KBr) 3700-2750 (br), 1720-1650,1525,1455,1440,1400,1370, with HOAc, evaporated under reduced pressure, and purified by 1350, 1260,1170,1060 cm-'. Anal. Calcd for C38H&601$ c, silica gel flash chromatography (CHzCl2to 5% EtOH in CHzClz 58.60; H, 5.95; N, 10.79; 0, 24.65. Found C, 58.58; H, 5.96; N, gradient) to afford pure trimer 63 (298 mg, 67%)as a white solid 10.52; 0, 24.40. mp 183.5-186 OC (recrystallized from CHzClZ-E~O);'H NMR To a stirred solution of 2.1 M HCl in EtOAc (35 mL) was added (250 MHz, CDC13) 8.82 (d, 1 H), 8.58 (t, 2 H), 6.98-7.77 (m, 41 H), 5.13 (m, 13 H), 4.88 (m, 1 H), 4.47 (m, 1 H), 4.28 (m, 1 H), fiiely powdered trimer acid 56 (2.23 g, 2.86 mmol). The mixture 3.84 (m, 1 H), 3.66 ( s , 3 H), 3.09 (m, 5 H), 1.40 (s,9 H); IR (KBr) was diluted with EtOAc (15 mL), stirred at room temperature for 1 h, diluted with dry EbO (150 mL), and centrifuged. The 3330,3100,3080,3040,2940,2880,1745,1710,1685,1660,1655, 1580,1520,1455,1375,1310,1265,1210,1170,1130,1090,1030 pellet was washed with EbO (5 X 25 mL) and dried to afford amino acid salt 57 (1.80 g, 88%) as a white powder: mp 211.5-213 cm-'. Anal. Calcd for C78H,8N6015:C, 69.94; H, 5.87; N, 6.27. OC (recrysfrom MeOH/EhO); 'H N M R (60MHz, CD,OD) Found C, 69.92; H, 6.03; N, 6.26. 7.2 (s, 15 H), 5.05 (8, 2 H), 5.0 (s, 4 H), 4.7 (exchangeable H's), Linear Enterobactin Analogue 66. A solution of trimer 63 4.1-4.6 (m, 3 HI, 3.2-3.8 (m, 6 H); IR (KBr) 3700-2300,3300,3030, (230 mg, 0.172 mmol) in dry formic acid (5 mL) was stirred at 2940,1720,1690,1660,1520,1445,1385,1290,1260,1230,1180, room temperature for 2 h, evaporated in vacuo, and triturated 1020 an-'.Anal. Calcd for C s 3 H ~ 6 0 1 0 CC, k 55.42;H, 5.50; N, with anhydrous EgO to afford formate salt 64 (201 mg, 90.970) 11.75; C1, 4.96. Found: C, 55.20; H, 5.40; N, 11.52; C1, 5.23. as a white foam: 'H NMR (250 MHz, CDCl,) 8.75 (t, 2 H), 8.63 Cyclo(tri-@-alanine)(60). (A) Diphenylphosphoryl Azide (d, 1 H), 6.98-7.70 (m, 41 H), 5.12 (m, 12 H), 4.82 (m, 1 H), 4.52 Closure. A solution of tri-@-alaninehydrochloride (59,150 mg, (m, 1 H), 4.17 (m, 1 H), 3.69 (m, 1 H), 3.66 (8, 3 H), 3.42 (m, 1 0.561 "01) in DMF (110 mL) was evaporated to 80 mL in vacuo H), 3.32 (m, 2 H), 2.72 (m, 1 H), 2.64 (m, 1H); IR (CHC1,) 3340, and cooled to -40 "C, and diphenylphosphorylazide (145 pL, 0.673 3060,3000,2940,2870,1730,1715,1680,1650,1570,1515,1495,

(recrystallizedfrom EtOH); 'H NMR (60MHz, MeaO-ds) 8.4-8.7 (br s , 4 H), 7.6-8.4 (m, 7 HI, 7.4 (s,5 HI, 5.443,2 H), 5.2 (s,2 HI, 4.8 (m, 1 H), 3.4 (m, 2 H); IR (KBr) 3380,3030,2920,1735,1720, 1660,1585,1520,1450,1425,1320,1290,1200cm-'. Anal. Calcd for C&H~N20,$'3: C, 58.74; H, 4.05; F, 9.96; N, 4.89. Found C, 58.44; H, 4.27; F, 10.10; N, 5.08. Dimer Amine Salt 54. To a solution of acid monomer 50 (1.92 g, 5.67 mmol) and HOBt (1.07 g, 6.97 mmol) in DMF (8 mL) at 0 OC was added DCC (1.11 mL, 5.82 "01) dropwise. The mixture was stirred at 0 OC for 1 h and at ambient temperature for 1 h, and then a solution of amine salt 52 (3.20 g, 5.58 mmol) and N-methylmorpholine (0.62 mL, 5.62 mmol) in DMF (8 mL) was added. After being stirred 20 h, the mixture was warmed to dissolve precipitated dimer; DCU was filtered from the warm solution and washed with DMF. The filtrate was evaporated to onetenth volume and purified by extraction [EtOAc/l M aqueous citric acid, NaHCO,(sat, aq) and NaCl(sat, aq) washes], yielding dimer 53 as a pale yellow solid in quantitative yield (4.50 8). Crude dimer 53 (4.50 g, 5.58 mmol) was dissolved in 7:3 CF3COOH-HzO (40 mL) and stirred at ambient temperature for 2.5 h Evaporation of the solvent and reevaporation of the residue from toluene (three times), trituration with EtzO, and recrystallization from EtOH afforded dimer amine salt 54 (3.84 g, 87%) as a pale yellow solid mp 157-161 "C;'H NMR (SoMHz, CDC13 + MeaO-ds) 8.3 (m, 4 H), 7.9 (m, 3 H), 7.4 (8, 10 H), 5.4 (8,2 H), 5.2 (s, 4 H), 4.5 (m, 2 H), 3.7 (m, 2 H), 3.3 (m, 2 H); IR (KBr)

3590

J. Org. Chem. 1981,46, 3590-3592

1450, 1370, 1310, 1260, 1210, 1140, 1080, 1040, 1025 cm-’. Formate salt 64 (51 mg, 0.040 mmol) in pyridine (1 mL) and acetic anhydride (1mL) was stirred at room temperature for 2 h. Purification by extraction [EtOAc/l N HCl(aq), NaHC03(sat, aq) and NaCl(sat, aq) washes] and silica gel column chromatography (7:93 EtOH-CHzClz) afforded N-acetyl trimer 65 (37 mg, 72%) as a white solid: ‘H NMR (250 MHz, CDCl,) 8.70 (d, 1 H), 8.49 (d, 2 H), 6.9-7.7 (m, 41 H), 6.11 (t, 1 H), 5.10 (m, 12 H), 4.79 (m,lH),4.44 (m,lH),4.31 (m, 1H), 3.63 (s, 3 H), 3.2-3.6 (m, 6 H), 1.47 (s,3 H); IR (CHCl,) 3340,3060,3030, 2930,2870, 1735, 1650, 1570,1510, 1450, 1370, 1310, 1260, 1210 cm-’. A suspension of trimer 65 (37 mg, 0.0289 mmol) and 10% Pd on carbon (8 mg) in 1:l CHzCl2-MeOH (4 mL) was flushed with hydrogen, and 2.38 N HC1 in MeOH (18 pL, 0.043 mmol) was added. The mixture was stirred at room temperature under an atmosphere of hydrogen for 12 h, and the catalyst was filtered (Celite) and washed with EtOAc. The combined filtrates were washed with 1N HCl(aq) and 0.2 M aqueous citrate buffer (pH 5.5). The organic phase was dried (NaZSO4),evaporated under reduced pressure, and recrystallized from EtOAc/hexane to afford linear trimer 66 (16 mg, 75%) as a white powder: mp 155-157 “C (precipitated from MeOH with HzO and lyophilized);‘H NMR (250 MHz, CD30D) 7.18-7.27 (m, 3 H), 6.88-6.94 (m, 3 H), 6.64-6.74 (m, 3 H), 4.75 (m, 1 H), 4.60 (m, 2 H), 3.86 (m, 1 H), 3.72 (8, 3 H), 3.50-3.65 (m, 5 H), 1.87 (s,3 H); IR (KBr) 3700-2500 (br), 1730, 1635, 1580, 1540, 1525, 1450, 1360, 1330, 1260, 1170 cm-’; mass spectrum (field desorption), m/e 740 (M+). A-Cis Fe(II1) Complex 67. Linear trimer 66 (0.9 mg, 1.22 pmol) in MeOH (3.0 mL) was mixed with 4 X lo4 M FeC1, in MeOH (3.0 mL). The resulting violet solution was diluted 1:l with 0.1 M aqueous phosphate buffer (pH 7.2) to afford a wine-red solution of ferric complex 6 7 visible spectrum (31pH 7.2 aqueous 495 nm (c 4400); CD spectrum, phosphate buffer-MeOH) ,A, vide Figure 1. Equal volumes of ferric complex 67 (1 X lo4 M) in 50% aqueous MeOH and ethylenediaminetetraacetic acid (EDTA) disodium salt in 0.1 M aqueous phosphate buffer (pH 7.2) were mixed, and the absorption at 495 nm was measured after equilibration of chelated Fe(II1) was observed (constant optical density at 495 nm). The absorbance at 495 nm was 4%, 9%, and 86% diminished at and lo-’ M EDTA, respectively.

By use of these data and a pKa of 10.1 f 0.2 for linear analogue 66, the formation constant was calculated; Kf = 1.2 (Kffor Fe(III).EDTA complex = 1025).46

Acknowledgment is made to the National Institutes of Health (Grant No. AM 20452 and GM ZSSZS),Eli Lilly and Co., and Firmenich and Co. for their generous financial support. We thank Professor J. B. Neilands (Berkeley) for biological testing and for his continued collaboration, Mr. Curtis Adams, Ms. Barbara Beiber, and Mr. Mitchell Weitz for technical assistance, and Dr. Catherine Costello for field-desorptionand exact mass spectra. We also thank Professors D. Kemp and S. Masamune for valuable discussions. Registry No. 5, 4726-96-9; 6, 69146-58-3; 7, 78088-92-3; 8, 40351-09-5; 9,78088-93-4;10, 78088-94-5; 11, 7724-78-9; 12, 7808895-6; 13,78148-29-5;14,3041415-4;15,78148-30-8; 16,3262-72-4;17, 78088-96-7; 18, 23680-31-1; 19, 78088-97-8; 20, 78088-98-9; 20 dicyclohexylammonium salt, 7808899-0; 24,78089-00-6;25,78089-01-7; 26,78089-02-8;27 (isomer l),78089-03-9; 27 (isomer 2), 78089-04-0; 28 (isomer l), 78089-05-1; 28 (isomer 2), 78089-06-2; 29,78089-07-3; 30,78089-08-4; 31,78089-09-5; 33,6081-61-4;34,75299-15-9; 34 THP ether, 78089-10-8; 3 (isomer l),78148-31-9;35 (isomer 2), 78148-32-0; 36, 75299-17-1; 36 THP ether, 78089-11-9; 37, 75299-18-2; 37 THP ether, 78089-12-0;38,75299-19-3;39,75363-09-6;40,75363-10-9;41, 75299-20-6; 42, 75363-11-0; 44, 16947-86-7; 45, 78089-13-1; 45 dicyclohexylammonium salt, 78089-142; 46,78089-15-3;47,78089-16-4; 48, 78089-17-5; 49, 7814833-1; 50, 16947-84-5; 51, 78089-18-6; 52, 78089-20-0; 53, 78089-21-1; 54, 78089-23-3; 55, 78089-24-4; 56, 78089-25-5; 57, 78089-26-6; 59, 78089-27-7; 60, 10491-78-8; 61, 78089-28-8; 62, 78089-30-2; 63, 78089-31-3; 64, 78089-33-5; 65, 78089-34-6; 66, 78089-35-7; 67, 78090-01-4; 2,3-dihydroxybenzaldehyde, 24677-78-9;benzyl chloride, 100-44-7;2,3-bis(benzyloxy)benzaldehyde, 5779-91-9; 2,3-bis(benzyloxy)boic acid, 74272-789; oxalyl chloride, 79-37-8;p-bromophenacyl bromide, 99-73-0;24bromomethyl)anthraquinone, 7598-10-9;p-bromobenzyl bromide, 58915-1. (46) OBrien, I. G.; Cox, G. B.; Gibson, F. Biochim. Biophys. Acta 1971,237,537.

Convenient Synthesis of the Pseudoguaiane Ring System J. Joseph Christie, Thankamma E. Varkey, and John A. Whittle* Department of Chemistry, Lamar University, Beaumont, Texas 77710 Received February 20,1981 Ketone 8 was conveniently synthesized from acyclic precursors in a four-step procedure. Condensation of which was alkylated with propargyl cis-1,4dichlor+2-butene with diethyl ketone gave 2,7-dimethyl-4cycloheptenone, bromide. The triple bond was hydrated with mercuric ion impregnated Dowex 50 and the compound cyclized with potassium tert-butoxide to yield ketone 8.

The pseudoguaiane skeleton 1 contains one of the more

1

2

common ring systems found in hydroazulenic sesquiterpen0ids.l Many early syntheses of the hydroazulenic ring system involved “special” reactions such as rear(1)See: Devon, T. K; Scott, A. I. “Handbook of Naturally Occurring Compounds”; Academic Press: New York, 1972; Vol. 11, pp 61-72.

rangements, ring expansions and photolyses. Over the past decade, however, many syntheses using “conventional” procedures have appeared.2 Recently, Lambury3 has used the different reactivities of the carbonyls in diketone 2, (2) See, for example: Marshall,J. A.; Partridge, J. J. Tetrahedron 1969,25, 2159. Kato, M.;Kosugi, H.; Yoshikoshi, A. Chem. Commun. 1970,185. Hendrickson, J. B.; Boeckman, R. K.,Jr. J.Am. Chem. SOC. 1971, 93, 1307. Heathcock, C. H.; Ratcliffe, R.; Van, J. J. Org. Chem. 1972, 37, 1976. Marshall, J. A.; Ruden, R. A. Ibid. 1971, 36, 2569. Marshall, J. A.; Green, A. E. Ibid. 1972,37,982. Anderson, N. H.; Uh, H . 4 . Synthe. Commun. 1973,3, 115. Anderson, N. H.; Uh, HA. Tetradron Lett. 1973,2079. Buchanan, G. L.; Young,G. A. R. J.Chem. Soc., Perkin Trans. 1, 1974,2404. Marshall, J. A.; Ruth, J. A. J. Org. Chem. 1974, 39, 1971. De Clercq, P.; Vandewalle, M. Ibid. 1977, 42, 3447. (3) Lansbury, P. T.; Serelia, A. K. Tetrahedron Lett. 1978, 1909.

0022-326318111946-3590$01.25/0 0 1981 American Chemical Society